I’m a bit upset, because I can come up with an investment idea and build personal exposure faster than I can write up a full report and build charts, which as you can guess, is a long and laborious process.

It’s a labor of love, and for the past 2 months I’ve been publishing reports weekly for free, with no paywall.

Which also means for the past 2 months, everything I've published for free has reached you after I already had the position on. That's the honest reality of a weekly newsletter.

But that changes on Saturday.

The premium version of Babylon Burns launches this weekend. My macro analysis stays in front of the paywall. Everything that actually makes money (the stock picks, the timing, the sizing, the positioning) goes behind it including numerous exclusive features.

Annual members get:

• Live trade alerts. You'll know when I put a position on, as I put it on, rather than waiting for the write-up.

• The full live portfolio. Every position, every stop-loss, every target, with the rationale for each.

• The companion web app, with my complete chart library organized by theme and every position tracked in real time. Mobile version to follow.

• All premium research going forward, including single-name write-ups you won't find anywhere else.

Monthly members get the full research, every premium write-up as it publishes. But the live portfolio, the trade alerts and the app are annual only. These letters aren't built to be followed month to month, and the pricing reflects that.

Now’s the best time to subscribe before we put the paywall up. Get started today.

I’ll be wrong quite a bit, and I’ll be right quite a bit, and being able to turn those insights into alpha is something that will require the annual membership.

The pitch:

In the autumn of 2025, one hundred and sixty six people took part in a study that would decide the fate of a drug.

Their task was simple: Open a small foil pouch, take out a strip of film about the size of a postage stamp, and place it under the tongue.

Nineteen of them put it on top of their tongue instead

That was enough for the FDA to reject the drug on January 30 of this year. Now, you should note that the agency raised no question about efficacy, none about safety, none about how the product is made.

The rejection was based on the fact that the pouch that took seventeen seconds to open and that 11% of people put the film on top of their tongue rather than underneath it.

Within three weeks the company behind it had lost roughly half its market value, because a rejection is a rejection in a headline.

Often with these small names, the market can’t always tell the difference between a drug that failed and a simple packaging problem.

Thankfully, I could.

The company is Aquestive Therapeutics, ($AQST), and I tweeted the call on July 29 at $3.60.

The stock has gained 20% since. I expect another 50% upside, a move to at least $6.50, by next spring, when I expect the drug to be approved.

And just this Monday, we got the first hard evidence that it’s on track: the redesigned pouch now opens in three seconds, the placement errors are down by more than 80%, and the study everyone feared came back very, very clean.

I’m laying out the full case in this report.

A good drug is one people actually use

Anaphylaxis is a sudden, severe, and potentially life-threatening allergic reaction.

It can be caused by peanuts, shellfish, a bee sting, and within minutes the person’s airway swells, blood pressure collapses, and without treatment… people die.

There is just one drug that reverses it: epinephrine.

It also goes by the better known name of adrenaline. It has no substitute.

The entire question in this market is one of delivery: how do you get this drug into a terrified human being rapidly. Self-administration, or a parent helping a child are the two common scenarios.

For half a century the answer has been an injectable needle. The EpiPen and its generics from Viatris and Teva own the prescribing habit; Auvi-Q is a talking auto-injector; Symjepi is a prefilled syringe.

They all work, and they all share the same flaw, which is that people don’t use them.

Patients delay because they’re frightened of the needle. Worse, they don’t carry the thing at all, because an auto-injector is a bulky plastic tube that doesn’t fit in a pocket.

Study after study shows the same pattern: fatal and near-fatal reactions where the device simply wasn’t within reach, or wasn’t used in time.

In August 2024, ARS Pharmaceuticals broke the needle barrier with neffy, an epinephrine nasal spray.

It proved the point the whole industry suspected: give people a needle-free option and prescriptions follow. neffy did $72 million of US revenue in its first full year, and CVS Caremark, one of the three big pharmacy gatekeepers, added it to formulary this July.

Anaphylm, which is AQST’s drug, is the next step in that logic, and I think it’s a step change in convenience and usability.

It is delivered on a film the size of a postage stamp that dissolves on contact under the tongue.

It weighs less than an ounce and you can pop it in your phone case or your wallet, which makes forgetting it less likely.

If the category’s core failure mode is that the medicine gets left at home, the small stamp is the best solution for that problem so far.

It would be the first and only oral epinephrine, and the FDA’s own review didn’t question any of its pharmacology.

Needless to say, this is a multi billion dollar market which AQST would be chipping away.

The company behind it

Aquestive isn’t a story stock invented for this product, which is refreshing, as it gives us a floor under the share price which is a setup which is easier to get behind.

It’s a twenty year old drug delivery manufacturer in, New Jersey, whose specialty is exactly one thing: turning drugs into stamp sized films.

Its base business makes Suboxone film, the opioid dependence treatment, for Indivior, along with a handful of smaller licensed products. That business did $44.5 million of revenue in 2025 and runs from a plant which the company owns.

This business is probably worth about $1.50 per share without Anaphylm.

Add in the option of eventual approval through a third or fourth submission, and failure from here looks like a $2 to a $2.50 stock.

What the FDA said, and the case against this idea.

A Complete Response Letter is the FDA’s formal no: the application won’t be approved in its present form, and here’s why.

Aquestive’s letter, received January 30, basically said that the label needed to be improved to address the film placement, the pouch was difficult to open, and asked for one supportive pharmacokinetic study.

According to the company, there were no clinical deficiencies, no safety deficiencies, and no manufacturing deficiencies. The drug’s core data, the head to head pharmacology against auto-injectors across eleven studies and 967 administrations, went unchallenged.

This was the whole case for it which got me interested, but before we go deeper, there are some clear counterpoints which I think it would be best to consider before we go down this road.

First, we haven’t read the letter. Nobody outside the company and the agency has.

Everything above is management’s characterization, and the FDA’s researchers once compared 61 of these letters against the press releases describing them: only 14% of the statements matched, and sponsors omitted roughly 85% of the agency’s safety and efficacy concerns.

Aquestive’s framing has been consistent for six months and is corroborated by the specifics it volunteered, but let’s not forget that we are asked to trust the party with the most to lose.

Second, Aquestive is second in its category. neffy is approved, funded, eighteen months ahead, and defended by a sales machine which Aquestive won’t have the resources to match.

And finally there is about $95mn of committed capital from Oaktree and RXT, which only comes through if the drug is approved.

So this isn’t an obvious layup operationally. I’m more interested in the binary nature of the approval though.

From here, I paused, and asked: what happened in other cases when companies got a rejection similar to this one?

This is where my detective work started.

Twenty one rejections like this one

There’s no published approval rate for rejections of this type, and the academic work that classifies rejection letters is too generic to be of use.

So I built the dataset myself, by hand, from FDA review documents and two decades of SEC filings, and then ran the entire EDGAR full text corpus across over 4,500 filings, to make sure I hadn’t missed cases.

I found 21 original applications since 2010 where the FDA’s objection was substantially about human factors, usability, device performance, packaging, instructions, or labelling.

Let’s look at some of the case studies, as it will give us a better feel for what we’re dealing with. You can speed read this bit if you need to.

First the good ones:

In November 2014, Avanir’s Onzetra Xsail, a migraine drug delivered through a novel breath powered nasal device, was rejected. The FDA’s summary review says the letter was issued “because of human factors deficiencies”: in the usability study, only a fraction of patients delivered a full dose correctly. There were no clinical questions, nor manufacturing questions. Avanir fixed the device, reran the study, and the same reviewer wrote that the sponsor “has adequately addressed the device usability issues” and approved it.

It took 14 months from the letter of refusal to the approval.

Kloxxado, an emergency nasal spray for opioid overdose, drew a 2020 letter in which human factors was one deficiency among twenty; the agency found “insufficient information to support the design of your proposed user interface.” There was one resubmission, and after 14 months, it was also approved.

Gimoti, a nasal spray, took a letter on device quality and clinical pharmacology with no new clinical data demanded: it was approved in 15 months.

Kynmobi, an apomorphine film for Parkinson’s is worth reading closely because its letter was worse than Anaphylm’s. The FDA cited human factors and inadequate bridging to the reference drug and inadequate characterization of adverse events, a mixed rejection that reached clinical ground. It still cleared on a single resubmission in 16 months.

MydCombi, an ophthalmic spray whose only sin was device documentation, took 18 months.

Now, the not so good ones:

Adamis’s Symjepi, an epinephrine prefilled syringe, is the case that shows us what the FDA sounds like when it doesn’t approve of the device: the agency wrote that the device as designed “could fail in the hands of the intended users,” and that in an anaphylaxis emergency, delivery of epinephrine could not be assured.

It took the company two rejections, a redesign and 27 months to approval.

Primatene Mist spent 54 months and two letters proving that ordinary consumers could read its label.

scPharmaceuticals’ Furoscix, a wearable furosemide pump, took 52 months and three review cycles even though the FDA twice said explicitly that no new clinical trials were needed, because the fix required modifying hardware.

Amphastar’s naloxone nasal spray, took 72 months, because the human factors remedy forced a device change, the device change altered the delivered dose, and the altered dose demanded new pharmacokinetic studies.

Teva’s generic EpiPen needed 30 months. Mannkind’s Afrezza inhaler, 52, after the agency demanded two new clinical trials to bridge a redesigned inhaler.

And 5 never made it at all.

20 of the 21 are resolved. 15 were approved, 5 died. 75% made it eventually, against a baseline historical base rate of about 73% for rejected applications and a 50% first try rate, from a decade worth of studies.

But the aggregates aren’t so interesting to me, it’s more the structure of the letters which gave us insight into when it would go

Every case that cleared on a single resubmission was approved in 12 to 18 months, with a median of 14.

Every case that needed more than one resubmission took 26 to 72 months, with a median of 44 .

There is a massive fork in the road which lies between the group that tried on the first fix, and those that didn.t

It’s whether the first fix works. When the remedy is packaging, instructions, and one repeat study, you get 14 months and done. When the remedy turns out to require changing the product, you get years, or never.

Where it gets interesting, in six of the resolved cases, human factors or packaging was the only problem, with clinical and manufacturing untouched. All six were approved. Anaphylm, per everything disclosed, is the seventh member of that group.

Plot Anaphylm on this distribution and the projection writes itself. The letter is dated January 30, 2026.

The single resubmission median of 14 months lands in April 2027, and the mechanics agree: a 3rd quarter resubmission plus the FDA’s standard six month clock for this class of response puts the decision between late March and June.

That’s the spring I’m pointing at, and everything now rests on one question, whether Anaphylm’s first fix works. Hold that question, it was mostly answered on Monday.

On Monday morning, August 10, Aquestive published the results of the redesigned program, and they read like the fast lane of the reference class.

The pouch that took a median 17 seconds to open now takes 3. Incorrect film placement, the error that triggered the entire rejection, fell by more than 80% against the failed study. Self administered pharmacokinetics matched clinician administered. And the study I flagged as the single biggest risk in this situation, the deliberate misuse arm testing what happens when the film is placed on top of the tongue anyway, showed clinically meaningful pharmacodynamic response.

That last result means even the residual error mode, now down to roughly one user in fifty, has acceptable pharmacology behind it. The fix, if we believe the evidence, works.

The resubmission stays on track for this quarter, and Canada, the UK and Europe are advancing in parallel, with European regulators already on record that no further trials are needed to file.

Back of the napkin valuation

If approved, I expect the stock around $6.50. It could be $7, it could be $8, it could be higher. Let’s be conservative and call it $6.50.

From September 2025 through January, while the market treated approval as likely, this stock traded between $6.00 and $7.55, on a weaker balance sheet than it will have at approval, with every revenue encumbrance already disclosed.

I see no reason the market stops short of a price it already paid when the question was still open, so I think I’m being very conservative here.

If rejected again, we’re probably looking at $2.25, as I suggested earlier. That’s the manufacturing business, the optionality of resubmission, and the remaining cash, net of debt, and it’s consistent with the $2.93 panic print in February, with now extra newfound negativity.

I put approval chances at 80 %. It has all the hallmarks of the fast to fix drug category which I just dissected, and it seems, none of the curve balls which delay it.

At 80 % the expected value is $5.65 using my conservative upside scenario.

Yet the stock today at $4.29 is basically still pricing this as a slightly below 50% chance of success.

AQST is a $540mn market cap. It’s quite frequent in these cases for the market to ignore the situation, and leave some free money on the table. I already picked up some, but there is still plenty, and it now more free than ever.

neffy’s expensive lesson, and who paid for it

The strongest objection to the investment case isn’t regulatory, it’s commercial: the main competitor, neffy, has had a launch with brutal economics.

ARS spent roughly $230 million on selling and administration in 2025 to book $72 million of neffy revenue, $3 out for every buck coming in, with a team of 148 reps, a national ad campaign, and a pediatric co-promotion reaching 20,000 clinicians.

If that’s what this market costs, a company Aquestive’s size just can’t afford it, and the billion dollar product management likes to talk about isn’t anywhere close.

But here’s the thing: neffy wasn’t just selling a spray, it was building a category.

Those dollars went to convincing allergists that needle-free rescue epinephrine is legitimate medicine, convincing patients it exists, and forcing the pharmacy gatekeepers to create a formulary lane that didn’t exist, which finally opened at Caremark this July.

That work is done, and it doesn’t have to be done twice.

Aquestive launches into the category neffy paid to create, with the more radical convenience story in a market whose core failure is that people don’t carry the product, a concentrated team of 50 to 75 reps aimed at the allergists who write most of the scripts, and its own film plant, so every strip is made at commodity cost in a factory it runs itself.

Conclusion

You’ll come to find that my work at Babylon Burns will span many asset classes and industries.

There will be macro opportunities and micro opportunities, like this one.

AQST could be worth $6.5 or $8. I can’t honestly tell you what the number is upon approval. What I can tell you, is that all of the fundamental work done on it suggests that approval is much more likely than the market is crediting.

The risk reward is very attractive.

On August 15th, this Saturday, to celebrate 55 years of Nixon debasing the US dollar and forcing me into a position where I have to go into the weeds to find opportunities for us to stay ahead, I will be launching the paid version of Babylon Burns.

For the past 2 months, I’ve given everyone a free ride to demonstrate the breadth of my work.

Saturday that changes.

If you pledge today, you’ll get access to:

• All of the stock ideas like this one, which will go behind the paywall.

• My live positioning and trade alerts in a portfolio which we’ll build from scratch.

• All of the above available in a members only web application (only for annual subscribers)

• And much, much more.

So use the button below to join us now, on what will invariably be an amazing adventure.

As Babylon Burns, we’ll light the cigars

PS: Joining now, means getting the discounted price. Post launch, the price of the annual subscription is going up, so now is the best time to get in on the action.

"Markets are never wrong; opinions often are." - Jesse Livermore

What I’ve learned from a decade and a half of daily obsession with financial markets, is that you can be right about almost everything and still lose money.

Conversely, you can be wrong way more than most would be willing to admit on X or Substack, and still win big.

From the very first stock I bought, (some piece of sh*t Canadian oil stock that increased 50% in two weeks on news that had nothing to do with my thesis), I was consumed by the idea of beating the markets.

I launched one of the most popular newsletters on SeekingAlpha helping individuals build dividend retirement portfolios. Our dividend portfolios have beaten the market with less volatility with widows-and-orphans stocks.

I run a hedge fund which has rotated through quant and discretionary strategies as the market conditions have moved. Our (KPMG audited) returns have trumped the market, although with more volatility.

It is my intent that I will continue to do quite ok with investing over the next few decades. Not because I never will have a losing stint. Not because I think my ideas are better than others. Not because I’ve figured out the secret sauce to special returns.

But rather because, as the late Yankees manager Casey Stengel would say “most ball games are lost, not won”.

If you can avoid losing, you can end up winning.

You see, I’ve come to believe that beating the markets is not unlike getting physically shredded: it is simple, but it isn’t easy.

Eat less calories than you expend, lift heavy objects, get 8 hours of sleep, drink lots of water.

Overcoming and transcending your base level humanity is a function of having a system, a discipline, and sticking to it.

This report will set out the rules of how I will build the live portfolio which will accompany the Babylon Burns publication when the paid version launches on August 15th.

Most of the “analysts” you’ll encounter online will never discuss these matters because they have never managed money professionally and they care more about being right than truly generating alpha.

(By the way it’s not too late yet to get in at the discounted pledge price before we launch.)

Subscribe now

I’m quite content giving the portfolio management system away for free. As Richard Dennis told Jack Schwager in an interview for his famous book:

I will be using more quotes than usual in this report, as I want to drive home that these concepts are not unique, they are mainly common sense which most investors just ignore.

Why not just index?

The results of index investing seem to be rather good when you look at history.

Warren Buffett suggested that for most investors, buying an ETF was the way to go.

And that might be true if the following conditions are true:

• You have no knowledge or interest in the markets.

• You are ok delegating your risk taking to the markets.

• You are content being merely average.

For large stints of time, be it from 1929 to 1940, 1966 to 1982, or 2000 to 2009, being an index investor has been an awful proposition.

As Mark Twain, one of America’s finest thinkers once said:

And you must realize that when Buffett recommended this in 1993, it was a variant view, with just 2% of assets in passive vehicles. When he reiterated this in 2002, it was once again a variant view with still just 10% of total assets in passive strategies.

Today, the majority of money is in passive strategies. If you include the number of active managers that hug the index and have no real views, then the number is even higher.

Being passive was novel in the 90s and 2000s. Today, at the very least, you should pause and reflect.

A market-cap weighted index like the S&P 500 works very well in prosperous times, and puts money on fire in difficult times.

There is no discussion of whether or not we are in an AI bubble. We are. The question isn’t when it ends, but what happens when it does.

There is no discussion of whether or not the US is in serious fiscal trouble. It is. The question isn’t whether it will debase its currency to try and fudge the numbers, only how it will do it.

There is no discussion of whether or not the world is fracturing into a multi-polar setup which we haven’t truly experienced in over 80 years. The question isn’t where the next conflict breaks out, but what it means for the supply chains we have come to depend upon.

You may believe whatever you want, but I believe that this is not an environment which is particularly favorable to passive funds over the next decade or two.

You must ask yourself, do you think there is more chance that the next decade looks more like 1966-1982 and 2000-2009 or more like 2009-2020 and 1983-2000?

Indexing is putting everything on autopilot. It works fine in fair weather, but when a storm hits it pays to have a steady hand at the helm.

Not unlike the little red hen, if we want to do better, we’ll have to do it ourselves.

So where do we start?

One of the biggest fallacies of investors is to listen to what successful investors say they should do, rather than doing what successful investors actually do.

I dug into my history of investing and uncovered something which I expected directionally, although not so precisely:

20% of the stocks I bought accounted for 80% of the portfolio’s gains.

Buffett, Druckenmiller, Soros, all share the same footprint, sometimes even more dramatic and concentrated.

We all have a few good ideas. We all have many so-so ideas. We all have some very bad ideas.

Investors tend to believe that their primary job is to find good investments. Finding good investments is necessary, but it is not sufficient.

Your primary job is to build a portfolio with a shape which allows you to have many small losses or underperforming investments that never matter and a few large wins which you allow to grow with no limit.

Over any ten-year stretch, a handful of positions will produce almost all of your profit. You cannot know in advance which ones. What you can control, completely, is the other side: that nothing ever hurts you badly.

Buffett would say that rule number one is to “never lose money”.

Yes. But the way you do that is by losing a little often, and winning big sometimes.

Human nature goes the other way, we are tempted to stick to our guns, double down on our false narratives, as our egos get in the way and we want to be “right” more than we want to make money.

So we want to give ourselves a system that allows us to express the power laws that occur in investing.

I have 3 big lessons when building a portfolio.

Lesson 1: Lay a seedbed

William Goldman was a screenwriter, and a successful one. His insight applies just as much to investing as it does to writing a blockbuster movie script.

While only a few of our investments will generate superior returns, it is very hard at first to know which ones will.

So you cannot go around waiting for a jolt of lightning to give you insight into the 2 or 3 big ideas you should buy.

You need a discipline, a process, a way of getting into the good ideas to start with.

You should try out many positions, all while being willing to kill most within a few months.

The mechanism is simple. When something catches your interest, buy a starter position of half a percent to one percent of your portfolio. Its purpose is not profit.

Its purpose is attention: real money, even a little, forces you to actually watch a thing, read about it, and form a view, in a way a watchlist never does.

I like the metaphor of laying a seedbed. Only a few will become saplings, and of those even less will grow into beautiful trees.

Of course you don’t pick stocks randomly (although I’m half convinced some industry professionals do just that).

You need a solid process of research and ideation which must start with curiosity. I get ideas from unrelated discussions with my wife, from X, from stock screeners, from reading the FT, and sometimes I even have ideas of my own.

My philosophy is to leave no rock unturned. Yes I have certain views on macro, yes I have certain tilts on sectors. Yes I have certain tilts on narratives and possible scenarios.

I’m generally looking for investments where the tape, the theme and the tale are satisfying. I’m projecting the future 6-30 months out, and investing based on my world views over those time frames.

I need some conviction to start a position but not heaps.

You need to be willing to have strong views, held loosely.

Lesson 2: Love to lose money, hate to make money

Like Mark Spitznagel’s mentor, Everett Klipp (also known as the Babe Ruth of the CBOT) would say.

Now of course, we are in this to make money, but the most direct path to doing so is one which is roundabout psychologically.

If you love to lose money, you’re willing to crystallize small losses. If you hate to make money, you’re late to crystallize wins, and adding to your winning positions.

Peter Lynch used to say that:

If you can kill your ego and allow yourself to be wrong time and time again, and if you can kill your ego and not be too fast to sell your best ideas, then you will be setting yourself up for success.

The best equity investment I’ve made was Broadcom($AVGO) . I bought Broadcom at a split adjusted price of $43, then added at $47, $53, and $57 until it reached 7% of my portfolio.

My average cost is about $48 making the position a 9 bagger.

But there were two, very obvious mistakes I made: The first is, I sold many of the shares too early.

I’d exited over 70% of my position by the time it had 4xed.

The second is I stopped adding to the position despite absolute conviction in the trade.

On January 19th 2023, when i wrote that ChatGPT marked a turning point in the AI revolution, I said when the stock was at $57:

AVGO remains very attractively priced. The market is wrong about AVGO, and it is easy to buy, buy and buy, until you hit a full allocation for the position.

My mistake, was that this was such a big fat pitch, on which everything was aligning, that I should have made it 20% of the portfolio.

As Stanley Druckenmiller said reflecting on his time with Soros:

Which leads us to the next lesson:

Lesson 3: Get out of your own way

Every good investor has a sob story like mine on AVGO where they made money but not nearly enough.

You need a mechanism to allow yourself to double down on positions when you have conviction in the setup.

Ironically, the rule which has most empowered me to let a trade grow big is always knowing when I would get out.

In starter positions I often see an emerging narrative that I like with a chart that looks supportive, and will have a stop loss set at a level dictated by the chart.

As I add to a position when my conviction increases (from more information, refining the thesis, or confirmation from price action), I always update mentally where I would get out.

Generally there are 3 types of positions that I might take: the value trade and the momentum trade, and event based trades.

Sometimes the former turns into the latter, as was the case with AVGO, and my failure to identify that shift is what got me to sell a large slab too early.

In the value trade, you are buying the recovery of some fundamental dislocation you have identified, so you set trip wires which are closely tied to that thesis not playing out.

In the momentum trade you are looking for increased velocity so it’s more about the narrative and the guidance than the actual results.

Then there is the event based trade.

You buy something because you think something is going to happen.

Here, the rule is quite simple: If you buy an asset because you think something is going to happen, and that thing doesn’t happen, you sell the asset.

So you want some easy fundamental trip wires which will help you get out. You can also have trailing stops, or other lines in the sand you can use to get out.

Selling is one of the most difficult decisions in markets. When you don’t know what should get you to sell, I like to follow Paul Tudor Jones general rule of “nothing good happens below the 200 day moving average”.

Once you know when you’re going to get out, as long as what you believed would happen continues happening you can confidently add to a position. Your best positions should be added to with cash harvested from positions which aren’t working out, and from capital gains as the positions proves it works out.

You are allowed to set a cap on concentration to meet your risk profile. In dividend portfolios I would cap positions at 7% on cost, and let them ride to 10-15 before trimming.

In an alpha maximizing portfolio, which Babylon Burns portfolio will be, there will be no price cap.

Turning the lessons into a system

I must confess that a lot of rules were used so far to effectively say:

• Leave no rock unturned.

• Take lots of bets.

• Cut your losers, add to your winners.

• Know when to bail.

At any point in time the portfolio could look something like the above: a handful of small starter positions, a few slightly bigger ones, a few much bigger ones, and one or two very big positions, and some cash.

The portfolio will likely start from smaller positions in the 1% to 5% range, to truly mimic how I’d go about starting a new book from scratch.

Once again, I want you to understand that if you were expecting a magic pill, I won’t sell it because I don’t have it.

What I do expect, is that Babylon Burns will help you implement a simple but hard system.

Here are the rules the portfolio will follow:

The way it will work is quite simple.

On August 15th, when we launch Babylon Burns, everyone who signs up for an annual subscription will get access to our companion web application, which will include all positions which we will open on the following Monday. (monthly subscribers get the research only, when they’re ready to go for a year they can)

I will also be publishing following the launch the quickfire rationale for every position. It will include the stocks I’ve highlighted so far in my research on Babylon Burns, as well as many more ideas.

In the web application, everyone will have access to stop-loss targets, and a quick rationale of why it’s included in the portfolio along with stock charts organized per theme for you to consult.

When transactions are initiated, you’ll get an email notification. We will upgrade this to a mobile app in short order.

My intent with Babylon Burns is that with a subscription you will get access to:

• My global macro research which will be focused on whatever is of value. (so far we’ve covered oil, gold, tech, and the fiscal backdrop).

• Single stock write-ups of idiosyncratic opportunities you won’t see elsewhere (so far we’ve covered engine maintenance, healthcare, and oil rig opportunities.)

• A live portfolio with trade-alerts, targets, charts, and everything described above.

Like Muhammad Ali reminds us:

If we build our portfolios the right way, then I’m quite confident that…

…as Babylon burns, we’ll light the cigars.

PS: Another friendly reminder that it is not too late to pledge a subscription at a discounted pre-launch price by clicking on the button below.

In this article I will lay out exactly how it does it, show the evidence each part, and tell you how this manipulation will end.

I’m not alone with this suspicion. Over the past few weeks, discussion I’ve held with industry experts are pointing the same way.

My research has been supported by evidence brought to my attention by Chris Matherson, and various gentlemen going by the pseudonyms of JustDario, Gavin McCracken, and others to whom I owe a debt of gratitude.

The US government has a motive, it has the legal capability, and it has demonstrated the willingness to intervene.

I spent the first decade of my adult life shying away from conspiracy theories. Free markets were great, and those claiming manipulation were the crazies who struggled to generate sufficient PnL.

But there comes a time when you have to consider all the evidence at hand and make an informed decision. I hope by the end of this report, you’ll do the same.

Why would the US even bother?

The most common objection I get to my oil suppression thesis is that if the government is suppressing oil, they’re failing, because gasoline is $4.10.

This is a fact which even confounded President Trump after the signing of the now failed MoU with Iran just a month ago:

This level of public acknowledgment of ignorance of how gasoline is priced only shows to me that he isn’t the mastermind behind the suppression.

The price of gas was never target number one. Rather, US treasury yields and treasury inflation breakevens are the target.

The federal government has to roll about $9 trillion of debt in the next year on top of a deficit which is running at 7% of the GDP.

The issuance of this debt will be set by the marginal buyer of treasuries, and that marginal buyer will set the price based on expected inflation.

And within the inflation basket, energy is the only component which moves every week, has a deep liquid, forward-dated market attached to it.

And because energy inflation is upstream of inflation in the whole economy, investors, fed officials, and everyone involves uses the crude oil forward curve when setting out their inflation path.

They will look through high gasoline today, if the crude oil curve is deeply backwardated, as is the case today.

That’s why inflation breakevens at different tenors have been coming down despite the price of gas being up. My study of 5 year breakevens show that they move 0.5bp per 1% move in oil.

If you can slam crude oil prices, you can slam inflation expectations.

This gives the Fed the cover needed to not raise the rates.

The last thing the administration wants is a rate hike before the November elections. Suppressing oil prices enables that objective.

It also justifies the war with Iran in a way which $150 oil would struggle to do.

That’s the motive: not letting power slip away.

How to actually manipulate the price of oil

There are basically three ways you can manipulate a price, and the taxonomy was established in research by Allen and Gale in 1992, where they concluded that all situations of price manipulation used a combination of the 3 ways.

• Information-based manipulation: using the power of narratives and releasing false information and spreading false rumors.

• Action-based manipulation: impacting the real world value of an asset through actions in the real world.

• Trade-based manipulation: accomplished through buying and selling without any public action or release of any information.

If you were going to build a sophisticated operation to suppress the price of crude oil, you would need to employ all 3 of these tactics together, as they are complementary and build off one another.

You would also need someone smart enough to orchestrate this who understands deeply how markets work.

You’ve got that person. He is the secretary of the US Treasury and his name is Scott Bessent.

I would argue that the US Treasury has never had someone at the helm who understood the game, and who was willing to play it to further the federal governments objectives and wellbeing. (note that this is a very different objective than that of furthering the general public’s wellbeing).

So with that in mind, let’s look at how the administration has engaged in all 3 types of manipulation.

Information-based: lies on TruthSocial

The first leg of a manipulation campaign has to start with (mis)information. It is humans, which through their bidding and asking set prices in the markets. These bids and asks are set based on expectations.

If you can change the expectations, you can change the price. A statement that cannot be checked moves a market only when the speaker's interests align with the listener's (Crawford and Sobel, 1982).

Most investors are overexposed to the AI trade, whether by decision or lack-thereof with passive investors piled into a trade they don’t understand. This is a very long duration trade, the type that doesn’t do so well when energy prices spike.

So the governments interests are aligned with Wall Street’s and Main Street’s interests.

But even those who aren’t so exposed, get affected. Public signals carry disproportionate weight due to the second order thinking associated: each participant trades not on what he privately thinks, but on what he believes everyone else will now think, and an official announcement is the natural coordination point for that second guess. A loud enough statement moves the price even among traders who privately doubt it, because each expects the others to move.

The oil market should be unusually resistant to this as flows are logged at customs, tanks are measured by satellites, and cargos are mostly tracked by their transponders.

But like Goebbels supposedly said if you “repeat a lie often enough and it becomes truth”.

I think it is a simple statement of the record to say that President Donald Trump’s TruthSocial feed has been full of falsehoods. Here is just a selected few.

Plus before that there were the dozens of posts between March and May saying that Iran wants the deal to happen, Iran has asked for a ceasefire, great progress is made towards a final agreement, a big day for world peace.

This “firehose” approach of rapid fire assertions with an indifference to consistency is a feature of the strategy. Repetition breeds familiarity, and familiarity gets mistaken for truth (Paul and Matthews, 2016).

The narrative was quick to supply an explanation for falling oil prices on the days it fell, and insisted on that narrative on days it didn’t fall.

Talk is cheap, and it should not survive contact with a well instrumented market. That is can and has are a consequence of the volatile market which we experienced in April 2025.

Market participants learned that following Trump’s “THIS IS A GREAT TIME TO BUY” post a few hours ahead of announcing the pause on his tariff campaign was a source of alpha.

The verification of the information becomes secondary, as market participants are coordinated by the president’s posts, regardless of the information value they contain.

Action-based: Draining the reserves

The information leg seeks to coordinate actors and confuse them on the elements that cannot be fully verified.

The second one, the action based leg seeks to reach into the physical world and change the balance of supply. Governments build up stockpiles of commodities, which they can use to manage the price of that commodity in a crunch. This is nothing new, and for oil, we’ve had the Strategic Petroleum Reserve for half a century.

Unfortunately the theory here is strong and carries a warning which those intervening cannot escape. Governments which defend a commodity price out of finite stockpiles create the conditions for its own defeat (Salant, 1983).

A stockpile which is large enough to hold the price is a a stockpile which must eventually run down, which every participant can see running down.

Once the defended price can no longer be defended, it snaps beyond its fundamentals. It’s like whack-a-mole. You can keep pushing it down, but sooner or later it must come back up.

In 1968, 8 central banks attempted to hold gold at $35, which lasted until the stockpile could no longer cover the claims against it, and then the market broke within days.

The closure of Hormuz knocked off 19.4mb/d. 6mb/d got rerouted through the Saudi East-West pipeline and the UAE’s Fujairah pipeline. That left 13.4mb/d. China, the US, and the other IEA particpants have been drawing down on their inventories and reserves to plug the gap, to the tune of 7m/bd.

That leaves a shortfall of 6.5mb/d, of which part is rationed by price (oil is after all more expensive than before the war), and part of which is a mix of undocumented Hormuz passages, previously sanctioned oil becoming unsanctioned, the East-West pipeline being pushed above its nameplate peak (not reflected in the chart above), and other unaccounted barrels.

Now this is fine and dandy, reserves are meant to plug and manipulate price in this exact situation, but here, we have reason to believe that it is done to paper over the fact that the market is actually extremely tight.

Consider that in October 2022 the Department of Energy announced a price band for crude oil at which it was interested in buying crude-oil: between $67 and $72.

Subsequently they rebuilt the SPR to 413mn barrels by the end of 2025 before being drained to 311mn, the lowest level since 1983 when the reserve was sill being filled for the first time.

Drawing down into an emergency makes sense. What it does not explain, is why, when the war ended on June 14th by the government’s own declaration, we continue to discharge a further 29mn barrels over the next 5 weeks, at a rate faster than during the war.

What’s more, these barrels were being sold by the end of the month at prices at which they had previously stated the SPR would be a buyer of oil.

Now from the 170mn barrels that the government was authorized to draw, 100 or so are gone. At the current rate of draw, we can continue until October. The IEA’s draws also.

The Chinese 5mb/d reduction in imports above was based on their 6.2mb/d imports in June. In July, these were increased to 7.8m/bd.

The Houthis have been attacking the Saudis, which is threatening the output of the East West Pipeline.

I will not engage in barrel counting and dating when things go wrong, because oil investors have collectively shown they misjudged a handful of events.

The only point that is important here is that the price has been suppressed by the US and other governments letting their stockpiles decline, and that this can go on maybe until the end of the quarter, end of the year if we got the back of the napkin math wrong again.

Physically plugging the market ensures that physical tightness doesn’t get to a point where shorts are squeezed.

Trade-based: shorting futures when nobody is looking

Manipulating paper trading is the third leg. It’s the one which has caught the attention of investors on X who are looking for proof of covert price suppression through flows.

It’s also the one which gets you put into the crazy bin.

I will present the evidence here which others have been more diligent in collecting than I have.

Price suppression in flows is by itself the weakest and most mean reverting of the 3 types of price manipulation and suppression. It works best when done in conjunction with the other two, as the price signal becomes the justification of the narrative, and the propped up physical market further corroborates this.

It creates a reflexive loop where the market inferred information impacts the real world and vice versa.

At least for a while.

Now before we look at whether the Treasury is shorting oil contracts, let’s first ask:

Does it have the capability to short oil futures?

And the answer is unquestionably YES.

The Exchange Stabilization Fund, created in 1934, allows the Treasury Secretary, with the President’s approval, to transact in securities and instruments of credit with no external oversight. It sits outside the appropriations process, and was used to fund the 1995 Mexican peso rescue because it routes around Congress, even when Congress refused to act.

It has been used after the Great Financial Crisis and after the Pandemic led crash. The vehicle for a covert market operation exists, and it has been used to bypass legislation before.

Now the availability of the mechanism is not a proof that it is being used.

The evidence which has been collected by fellow investors and analysts below is suggestive of certain actors attempting to suppress prices.

My takeaway is: if you are willing to drive inventories to the ground, willing to engage in a firehose of lies and misinformation, why not, if you have the ability to, also jam the signal with trade based manipulation?

Two gentlemen have done a very good job at scoping at what seems to potentially be the footprint of intervention in public markets. Chris Martenson identified abnormal volumes at 4.29am, where large amounts of Crude got shorted during weak market hours.

Another, who goes by the name of JustDario, has been highlighting the frequent campaign of short selling happening again and again in between 1.30AM and 6.00AM

The direction is definitely there. In the past 60 days crude oil has declined over night only to recover during the day, which is consistent with price being slammed during the illiquid thin hours only to recover thereafter.

Look, in my opinion, this is suspicious. It’s not the smoking gun that some of us oil bulls would want it to be, and I want to be cautious about the possibility for confirmation bias.

Nonetheless, I think it is safe to say:

• The price has been suppressed by using governmental and commercial buffers.

• The price has been suppressed by a firehose campaign of misinformation.

• This might even have been amplified by a covert trading operation.

While there might still be some debate on how oil prices are being suppressed, it should be extremely difficult for anyone to openly claim: we’re letting oil clear at market prices.

Such rigging of the prices must come with consequences. Some of these are already playing out, others are still yet to unfold.

Let’s take a closer look.

The pros have left the trade

The weekly CFTC report shows who holds crude contracts. Managed money, which includes the funds who usually go long into tightness and short slack in the market usually follow a typical, reliable pattern. When there is steep backwardation, they are pulled long into the trade.

Backwardation between the front and second contract has been particularly steep, and historically when it’s been this steep, managed money has been net long 270K contracts on average. Right now, they’re net long 87K contracts, which is the 15th percentile of the past 20 years.

Either the funds know something about the fundamentals which the general public does not, or they have learned that the price is being set by a group which doesn’t care about the fundamentals, and they’ve learned not to fight the tape and leave the trade all together.

Having market participants vacate the trade does the suppressors work, as an asset with less natural longs needs less selling to hold down.

The refiners are getting a free lunch.

I’ve been chanting this for a while, but if you suppress crude but not the products refined by it, then with tightness held equal, cracks widen quite naturally.

The 3-2-1 crack which measures the spread between 3 gallons of crude against 2 gallons of gasoline and 1 of diesel.

Refiners in the US have been running flat out at 96% to capture once in a generation cracks.

There is no spare refining capacity to be added on demand. Gulf refined products have been shut in, Russian and Saudi refineries have been blown up, Chinese refineries have reduced runs, all events which are downstream of conflict and locked in barrels.

Because crude is kept mechanically low, products continue to clear at market prices. A stressed but elevated $30 crack would suggest crude to be up to $40 higher than it now is, yet it isn’t because it’s suppressed.

Trump doesn’t understand this and Bessent doesn’t care. It’s why the price at the tank is still so high.

That is the tough part to stomach morally. The suppression which the government is undertaking is being done to support the forward curve, pin inflation expectations and yields, and further their objectives. It does nothing for the person who has to spend 35% more on a tank of gas than they did at the beginning of the year.

The tanks are draining

Suppression of the oil price into tightness with steep backwardation incentivises refiners to run hard, exporters to move US barrels onto the sea, and for inventories to be depleted, as buyers continue to buy at rates which are not sustainable due to reduced flow and supply globally.

The last time total inventories was this low was in 1984 and we haven’t finished draining them yet.

Cushing hit tank bottoms and has pretty much stabilized at the minimums. Commercial crude, gasoline and diesel all sit below their seasonal norms.

Drilling postponed

The slowest and probably most damaging cost, is that producers in the US, and particularly in the Permian are being sluggish in drilling new wells. The Dallas Fed Q1 survey put new-well breakevens in the $63-70 range, and we saw rig counts decline while spot hit $115, as capital discipline and a difficulty in pricing the duration of the conflict postponed decisions with the oil strip for next year sitting at their breakeven.

I wrote a whole piece on why I built a position in a micro-cap super-spec rig company to play what I see as the invariable snap once the charade is over.

You can read it here if you’re so inclined.

But this means that the second order effect of suppressing price is that it postpones Capex decisions which postpones when new supply will hit which only exacerbates the problem.

Governments have done this before, it always ends the same way.

Governments suppressing prices isn’t a new theme, it’s been tried in the past and it usually depends on one factor: whether the government can print what it sells.

As long as there are inventories to draw, it’s as if the government were printing barrels. The problem is that in this case it has a finite stockpile, and that it is also dependent on China, which is not an ally, continuing to suppress the price alongside it.

The US Treasury capped silver for decades and sold from its stockpile, which ran out in 1961 and forced silver to leave coinage within 4 years. There was the gold example from earlier which also ended badly in the late 60s.

Then there was the Megatons to Megawatts program which between 93 and 2013 turned Russian warheads into reactor fuel, an artificial supply that held uranium near $10 for two decades. At the end the repricing was violent because the Western countries had given up on adding supply into a market which was emptying a finite stock pile.

The lesson from history is unanimous: a finite defense against fundamentals buys time, transfers a fortune to whoever successfully times the break, and ends discontinuously.

If you know what’s good for you, own energy stocks

After the oil trade failed to truly catch fire in between April and June, I have been approaching the oil trade differently.

I still own a healthy amount of oil positions which I built up in 2020 and 2021 and never sold which were more investments.

This time around though, I’m thinking about it differently: what are the plays which are inevitable due to the physical constraints that emerge due to the conflict. The US rig play I mentioned was one. Down the line there will be the refinery maintenance play, although that is still too early.

Long term everything Atlantic basin should carry a risk premium to Middle Eastern barrels especially if the situation doesn’t resolve cleanly or fast (neither have been true so far).

There are pockets of the market which are interesting trades with a lot of potential torque.

What next?

When I launch the paid version of Babylon Burns on August 15th, annual subscribers will get my full list of energy investments in the companion web (and then mobile) application which will give you trade alerts, all my research, all my charts, and a portfolio so you can follow me along.

The post launch price will be higher than if you pledge today, which is why I highly encourage you to pledge right now.

I’ll be preparing extra bonuses for everyone who pledged early, as a token of appreciation.

We’re living in a complicated world. Fiscal largess, shifting demographics, saturated debt loads and productivity which doesn’t find its way into incomes have turned us away from the globalized status quo we came to expect.

The Middle Eastern conflict and the subsequent misinformation and price suppression campaign are only symptoms of a larger regime shift which will dictate the next 20 years.

Our only chance at outrunning it, is to generate returns which protect us in downturns, give us ample upside in bull runs, and keep us sane in markets which misbehave.

That’s what I’m striving for with Babylon Burns. In the coming world, you either do the getting or get got.

It is my intention that, as Babylon Burns… we’ll light the cigars.

Subscribe now

Since 1901, the United States federal government has collected more income than it has spent in only 31 years.

8 of those, came in a single stretch between 1920 and 1930.

Since 1960, there have been only 5 surplus years.

The last time was in 2001, and I was in primary school. In 2025, the government brought in $5.2tn and spent $7tn, a gap of $1.8tn in a year with no recession, no war and with unemployment at a very low level.

You must realize that this is now the default operating model of the US government and it will continue to be for the foreseeable future.

I wrote about why this won’t reverse in detail in my first article on Substack (“They sold you a dream and you bought it wholesale”).

But it comes down to this: For most of the postwar era, the expenses > income arrangement was tolerable because GDP growth carried the extra debt load that the deficit caused.

Gross federal debt peaked at 119% of GDP in 1946 and fell to 31% by 1981, A 35 year deleveraging which was achieved while still running net-deficits during the period.

But that escape has become less viable as real growth averaged 4% in the 50s and 60s, 3% in the 80s and 90s, and 2% since the financial crisis.

Meanwhile the debt ratio has crossed its WWII peak in 2020, and after a small reprieve, is once again knocking on the door of its ATH.

The last time the ratio was this high, it got there as a consequence of a war, and it took 3 decades of financial repression to bring it back down.

As you know, overextend sovereigns have 3 exits: default, austerity, and debasement. I’ve made the argument in the past that there will be more debasement, as default isn’t an option for the US and there is no political will for austerity.

Today we will be looking at a different angle: If there will be no austerity, who will be the beneficiaries, and how do we express that as a trade?

(This is a stock idea I presented on George Noble’s “best stock idea” conference this week, he is still selling the replay to it if you want to check out the other 14 speakers’ ideas too).

Subscribe now

Old people want their entitlements!

The Department of Government Efficiency launched 19 months ago in January 2025, with a stated target of identifying $2tn in spending cuts, with America’s richest man at the helm.

Its charter expired on the 4th of July, with only $215bn in savings claimed, which is about $0.11 on the dollar. Furthermore it’s a figure which couldn’t be verified and might have been optimistic. That’s not the point.

The point is: over the same period, Federal debt by several trillion dollars.

It was a doomed effort. 60% of Federal spending is mandatory, written into law, and a task force could never cancel these entitlements.

So even when there was political will to attempt government efficiency, the needle wasn't moved in any significant way, and there are no realistic paths to resolving this.

Why?

The answer lives in the constituency of voters in the US. In the 1964-1966 election cycle, voters over 65 cast 15% of the American ballots.

In 2020-2022, they cast 27.8% of the ballots.

It is quite easy to project where this is going in 20 years, as everyone who is already born is already born.

By the mid 2040s, over 65s will cast 1/3rd of the votes. Over 45s will be 2/3rds.

Singapore’s founding PM, Lee Kuan Yew saw this exact configuration and its problems coming over 3 decades ago.

In a 1994 Foreign Affairs interview, he argued that the one person, one vote was a British system inheritance which didn’t provide the optimal outcomes for a democracy.

He suggested that citizens between 40 and 60, which were raising families, should carry 2 votes each, on the grounds that a parent votes for the future as well as the present.

He assessed that elderly voters were troublesome, as an aging electorate would grow “tempted to pressure” governments for subsidies and assistance, regardless of the long-term cost to the state.

The United States is now well underway to living LKY’s nightmare. An ageing electorate will not vote for austerity, and politicians will have to be ever more considerate towards their needs, as their weight in the ballot-box continues to rise.

So where does the money go?

The most obvious destination for transfers to elderly people is healthcare.

Old people are like old airplanes ( I wrote an article on old airplanes), they need more maintenance.

Unsurprisingly over the past 25 years, as the percentage of the population over 65 has increased, so has the percentage of GDP which is spent on healthcare in the US.

Spending on care is mandated by law, as mentioned above, and it is compounding. It’s a simple line item where visible demography paths become easy fiscal arithmetic.

This leaves the state with one single lever: to provide the same care, but cheaper.

This is a lever that the federal government has been pulling for 4 decades. How do they achieve it?

It’s quite simple. Care in a hospital is more expensive than care at home. In 1981, home and community-based care was 1.1% of Medicaid’s long-term care.

This level crossed 50% in 2013, and currently is around 64%.

The fact that there is fiscal pressure due to the deficits will continue to accelerate this transition.

Care and benefits are promised to people. They continue to cost more. There will continue to be an effort to shift as much of this as possible to home-based and community-based care.

“But Sam, what about the cuts to Medicaid?”

The 2025 Budget Reconciliation Act cut $911bn from federal Medicaid financing over the next decade.

The response was to discount everything with Medicaid revenue, as if the cuts would fall evenly across the board.

A cut of this side (which will be within a total which keeps growing by the way) is a search for bloated fat. There are some clear rules concerning what can be cut: a focus on able-bodied optional adult services.

There is however, one subset of services which cannot be touched either legally or politically, and it lives at the other end of the spectrum of the oldies which we said will continue driving the overall number up.

Because that’s what people don’t understand about the cuts. They’re not cuts. They’re a reduction in projected growth. Federal Medicaid outlays in 2035 will still be meaningfully higher in nominal dollars than they are today.

And one area where they will not cut, is medically complex children.

And in particular, private duty nursing for Medicaid children.

Legally, the EPSDT mandate of 1989 obligates every state to cover all necessary treatment for eligible children. It’s a benefit which cannot be deleted.

Politically and morally, at a committee hearing, you’re not getting anyone vote to unplug our sick children while simultaneously increasing how much we’re spending on grandpa.

And given that states have to cover medically complex children, they would much further like that care to be administered at home, as it costs 10x less than in a hospital.

Which leads us straight into today’s stock pick.

(At this point after the launch on August 15th, You will see the paywall go up. But I’m giving everyone a free ride until then, if you enjoy the articles, please subscribe below to receive the next ones)

Subscribe now

Introducing Aveanna Health (AVAH)

Aveanna Health is the national platform for the exact niche discussed above: skilled nursing, in homes, for ventilator and technology dependent children, paid almost entirely by Medicaid.

It effectively sits at the intersection of all of the forces described above. It is one of the transfers the state cannot and will not stop paying.

The sorting process of what stays in Medicaid and what doesn’t always spares this demographic, at the very same time that the fiscal stress accelerates the substitution towards home care.

And as I will lay out below, you’ll come to realize that there are 3 sources of growth which the market has totally discounted in this name:

1. a caregiver labor pool which is finally loosening after a shortage which nearly bankrupted this business.

2. a contracted shift of volume towards premium rate payers that do not depend on further state reevaluation (even though, further state reevaluation is also likely)

3. and a transfer of enterprise value from the creditors to shareholders as the business continues to delever.

In a nutshell that’s the thesis.

Aveanna runs the same transaction, 46 million times per year: a nurse enters the home of a medically fragile child, the nurse costs a wage, the government program pays a rate, and the company keeps the spread.

In fiscal 2025, the rate was $43.39 per household, the cost was $30.56, leaving $12.83 of spread across 46.1mn hours of care.

So it is clear to see that Aveanna as a business has significant operational leverage with nursing costs and program rates competing in a tug of war for margin.

Now imagine that the 2017 private equity merger which created this company levered this thing 11x with debt.

When the post-pandemic wage inflation forced the cost line up 4.6% in 2022 vs frozen state rates which only increased at 2.6%, the spread compressed by $0.23, and hours stalled as the reimbursement couldn’t fund competitive wages. This was enough to tank EBITDA by 30%, and for the equity to fall by 94% from its $12 IPO to $0.67.

Management spent the next 2 years campaigning statehouses, and finally in fiscal 2025, the rates increased by 10%, the spread was up $2, and EBITDA shot up 75%.

The combination of operating and financial leverage gives this stock a beta of 1.9x and this isn’t going away.

The company’s revenue base is a fiscal transfer in pure form (with a modifier which I’ll explain later). Medicaid managed organizations supply 59% of revenue, direct state Medicaid is another 22%, Medicare is 10% (through a small adult home health segment which isn’t related to the thesis), and commercial insurance is 9%.

The business is spread over 38 states with concentration in Texas, Pennsylvania, California and following the June acquisition of Family First, a new fourth pillar in California.

Capex is $18mn per year against $2.6bn of revenue. This is an asset light business where the assets are the people, and the liabilities are still the private equity' era’s $1.48bn of floating debt rate.

The two levers

The recovery from the 2023 lows has ran on 2 distinct levers which are important to understand, as one of them will not provide the tailwinds which it has during the past couple of years.

The legislative lever is mostly exhausted here. Years of state-rate campaigns have produced double digit reimbursment increases, but as management informed us during the May earnings call, we should expect a stable rate environment with cost-of-living adjustments, with fewer state-rate wins expected.

In other words income will move with costs, and there won’t be a meaningful widening of the spread from here. In fact it is guided at $12.2 this year vs $12.83 last year.

The first quarter was the last year of the level change rerating, with 25% organic EBITDA growth vs only about 5% organic growth for the full year.

The second lever however, is still in play and offers a very attractive ramp in spreads.

Most Medicaid children are managed by capitated insurers, who receive a fixed payment per member and then eat the cost of care.

For that insurer’s economics, a child in a hospital bed is an economic nightmare. On the other hand, that same child staffed at home is a negligible expense.

But in a market where the number of available home nurses is the bottleneck, these payers are willing to pay more than the base Medicaid schedule to get a staffing priority.

Aveanna sells preferred payer contracts which give a staffing priority against rates 25% to 30% above the base schedule.

Now to be clear, the problem isn’t that there aren’t enough nurses, it’s that Aveanna is competing with what hospitals are paying for them. With hospital net hiring halved, and the hospital bid having collapsed after the pandemic induced squeeze, the marginal nurse is attainable to Aveanna.

Only a national platform like Aveanna can give a payer a staffing guarantee, which gives it a moat on which it can grow.

These preferred agreements covered 57% of the private duty managed care volume in Q4 2025, 60% in Q1 this year, and a management stated long-run destination to 80%.

This means that without including any volume growth, we can expect a contribution to income growth without any state intervention.

But here’s the thing: volume is UP and will continue to be.

Hours of care went from 37.9mn in 2021 then sequentially 38.5, 39.8, 41.6 and then 46.1mn in 2025.

With over 220,000 new RN licenses now minted annually against a hospital bid whose travel premium has fallen by more than a third, and Family First contributing roughly a million hours across its first seven months, I model high single digit organic volume in fiscal 2027 and 2028.

The risks to the model

Now all of this sounds good, but there is a risk to the model, and it depends on whether states decide to underfund the rate. While states cannot deny children the benefit, they can choose to underfund the rate. Cutting from where rates are currently is unlikely as it is politically explosive, but freezing can be rational for a while.

At a typical 60% federal match, cutting a dollar of nursing saves the state $0.4 today, while the 10:1 hospital tail costs amount to $4 later only if the unfilled hour of nursing care turns into institutional days down the line.

At the margin, the sad truth is that it usually turns into an unpaid mother staying at home instead, which weighs down the asymmetry in favor of freezing the rates.

The post pandemic cycle showed that the cycle went: freeze, crisis, correction, in that order.

If wave inflation resurges, the live political risk is that we once again see a freeze into rising costs squeeze Aveanna.

Thankfully, the current cycle is unusually well protected against both forms. Texas has no session until January 2027, its rates are appropriated law through August 2027, and its own Medicaid agency has quantified pediatric nursing underfunding at 21%, pre-drafting the case for the next increase.

Florida has a proposed 1.3 % squeeze on managed care capitation, not the schedule itself. This is marginal as payers are moving towards the preferred payer model anyway.

Pennsylvania's contractors report 27% of authorized nursing hours unfilled every month, the strongest rate-increase predicate in any state.

California just rejected its proposed home care cuts.

So I don’t see a big swing risk here, and the transition towards preferred payers further softens the impact of the states’ decisions here.

What is assumed by the current share price

So far we have a great story. And a great story is usually part of an investment.

The remaining two pillars are valuation and the technicals. Is it a good investment on paper? And does the tape agree?

My favored approach to valuation is to reverse engineer what must be assumed by the market for today’s price to be fair.

I found out donkey’s years ago in business school that DCF models were only as good as the inputs, and that the inputs usually weren’t any good.

At $9.09 per share, the equity is worth about $2.bn, and adding $1.4bn of net debt brings enterprise value to $3.4bn.

After converting EBITDA to FCF and making adjustments for stock compensation, capex, working capital, we get a business which yields about $210mn per year.

If we solve for the growth required that justifies the EV at a 9% cost of capital, then the market is requiring 3.6% annual growth for a decade, and 2.5% thereafter. At a 10% WACC it requires 5.6% growth.

But my back of the envelope math suggests that volume alone should contribute about 6 points of growth, the transition to preferred payer agreements should add another 0.7. State cost-of-living adjustments like contribute another 2 points.

This gets us to somewhere in the 8.5-9% growth. most of that does not depend on political risk and is mechanical.

And all of that is at the EV level, not the shareholder level, which is where equity investors sit.

At the 2022 low, shareholders owned 9% of this enterprise, with creditors due the rest.

Today the equity slice is up to 59%. By fiscal 2028 it will be 71%.

Even if EV didn’t grow (it would), the transfer from debt-holders to equity holders should rerate the stock to the tune of 20% over the coming years.

In a bear case where state rates get cut, margins erode, and the market requires a punitive 13.5% discount rate, the shares would be worth $6.98. In management’s base case the shares should be worth $13.5 today. In a bull case with prefered payer penetration reaching 80% faster than planned, shares would be worth about $19.

At $9.09 today, the market is over-indexing the bear case, when the baseline math suggests the situation is in fact much better.

The chart

There is good and bad in $AVAH’s chart.

As I might have made clear by now, I will be using technical analysis to validate entries into good narratives. Why? Because even if it is not perfect, that it is as much an art as a science, it is better than doing nothing.

We want to invest with the trend, we want to have rules that allow us to be wrong, we want to infer information from the buyers and sellers of the stock that give us an extra data point in making our decision.

The first obvious point with the AVAH chart is that it has been in a solid uptrend since 2024.

That trend is intact, and has been tested multiple times. The only time it was breached was a shake-out before the stock drove up 50% in a single week. So the trend is fundamentally on our side.

Where it gets a bit more complicated, is that despite the general movement upwards, it has now failed to break through the $10 level and stay there twice.

This creates a horizontal level at a psychologically important price which needs to be breached for the thesis to work out.

This means there is the potential for a sideways grind all the way through 2027 before the market realizes the upside which is rendered possible from the analysis above.

In these situations, we face a dilemma where we would actually prefer the stock above $10 than below as an entry.

The way I approach these scenarios, where I have enough base conviction in the idea but not in the entry, is to build a 1% position.

A note on portfolio construction

The 1% entry is a trick I learned from studying Druckenmiller’s portfolio allocation. Each quarter, he has 15-25% of the portfolio allocated to small new positions which didn’t exist in the prior quarter.

If you don’t have skin in the game, you don’t track positions well enough. There are too many narratives, too many stocks, something which hasn’t received a dollar won’t receive your attention.

Later, as the thesis develops and conviction grows you can double down, multiple times if needed.

That doesn’t mean all positions should be initiated at 1%. I’m comfortable to initiate anywhere between 1% and 6% based on conviction. If the chart isn’t supportive , it skews lower.

I will do a follow up piece, where I will explain clearly how the Babylon Burns portfolio will be constructed.

It is an aspect of investing which a lot of newsletters do not cover, because they are analysts rather than practitioners.

I’ve run the Dividend Freedom Tribe on SeekingAlpha’s portfolios for the past 6 years (all 3 of which have beaten the market on a defensive mandate), and I’m the CIO of a boutique hedge fund called Havelet Bay Capital (I need to check what I’m allowed to disclose, but our KPMG audited statements show we demolished the indices since inception).

I’m both an analyst and an investor. Babylon Burns will be an offensive portfolio. Not that I expect it will be riskier than investing in an index fund (most of which are offensive but people just don’t realize). But it will be conviction driven.

It will be a mix of trades, built on the principles which I’ve laid out in the other reports I’ve published and will continue to publish in coming weeks.

The losers will be cut, and the winners will be doubled down on.

Anyway back to AVAH.

When do we cut AVAH?

On these starter positions, you want to allow yourself to be wrong cheaply.

If we zoom into the weekly chart, and look at the triangle which AVAH broke out of in July, we can easily infer the point at which we’d want to bail.

This breakout fails if the stock price goes back below the prior resistance line, in theory. Some leeway must be permitted, and the trick to set that leeway is to look at the last fully contained candle, which I circled in yellow above.

This is a trick I learned from Peter Brandt (who you probably heard of if you read Market Wizards). You’d want to set the exit just below that candle.

This gives us a stop loss target of $7.6.

If you’re allocating 1% of your portfolio, a 17% loss is 0.17%, which is a rounding error.

And that’s the point. My work here is to find great narratives, uncover value by leaving no rock unturned, and then express that as conviction. When conviction is still building, we allow ourselves to be wrong cheap and fast.

I’d add to AVAH if it can close above $10 for a couple weeks.

What comes next?

On August 15th, I’m launching the paid version of Babylon burns.

Before that, I’ll be publishing 3 more totally FREE articles, including the portfolio building masterclass mentioned above.

When I launch the paid version, everyone who signs up for an annual membership will get access to all of the premium research going forward, and our member platform (which we’re working on building also as a mobile application), which will include the portfolio, the watchlist, all of my charts organized in libraries, and direct notifications when trades are triggered.

Below is a preview of the portfolio tracker, with the NorAm position initiated from last week’s article. (for demonstration purposes).

If you want to join Babylon Burns, now is a great time to pledge a subscription. Whatever I settle on as the final launch price will be guaranteed higher than what the pledges are set at right now.

I’ll write to you next week.

As Babylon Burns, we’ll light the cigars.

Oil spiked, then came down dramatically on the back of jawboning, a fog of war that extended into the media, and a control of the narrative which suggested that the ceasefire was a stepping stone towards durable peace.

My 2 main picks during the crisis, Petrobas ($PBR) and Frontline ($FRO) are up a meek 8% and 12.5% from when I said “buy”, which quite honestly is not enough compensation for the emotional rollercoaster we went through.

I held those positions and some calls on $BNO which still look like they will expire worthless. I chose to move my focus elsewhere and come back if I saw a setup which would revive the thesis.

The oil market didn’t respond as anyone would have expected.

Either the industry has stopped behaving rationally, or there is something that everyone is missing. Shale was supposed to be the world’s swing supply, but so far in this crisis, it hasn’t stepped up, with oil rigs still at historically low levels.

Nonetheless, I believe now, once again, the risk/return on the oil trade is in our favor, as long as we make it in a differentiated way.

This article will bring you through the logic of why I believe that a bet on rig count and daily rates in the Permian Basin is one of the most interesting setups right now.

I will also offer a stock pick which I believe is particularly interesting.

The only credible short term supply left is the Permian

We are all finding out in real time that the war is not resolved. Renewed strikes in the past week are likely cementing impairment in Hormuz which cannot forever be compensated with drawing of the SPR, in the US but also in China.

Incremental supply comes from only a short list of places: Guyana and Brazilian pre-salt will continue to provide more, but on a known schedule as there is nothing which can be decided today which will result in more barrels before the end of the decade at the earliest.

Washington has already used its trump card (no pun intended) of un-sanctioning barrels.

The only credible supply on earth that can turn a decision today into production within a year is the Permian.

And the Permian supply is a function of drilling rigs.

I believe that the current administration will continue to attempt to dunk on every rally in crude oil, and at times succeed.

As a lot of us found out in the first round of the conflict “we fought the law, and the law won”. China is and remains the wildcard that blindsided a lot of us.

But the price of crude seems to have settled above the $70 Brent on which Chevron built its 2030 Permian plan.

The longer the price stays above this level, the more likely we will see an increase in the number of rigs, at the very least, it puts a floor under the number.

Chevron is one of the most conservative planners in the space. At $70+ its 2030 plan is fully funded. Exxon, which is the other major which dominates the basin, has committed to growing Permian output.

The Dallas Fed Energy Survey is seeing year end price expectations being reanchored higher, as operators are slowly forming a view in line with what I’m saying.

Understanding the decline problem in shale

Conventional wells produce oil at roughly stable rates for decade. But horizontal shale wells produce a surge upon project completion, which then declines by 65% to 85% within the first twelve months.

The consequence is that shale basins are flows which must be continuously rebuilt. The industry measures this rate by what they call “base decline”: the share of production which will disappear within a year if drilling stops.

The Permian basin’s base decline was approximately 34% in 2018, 40% in 2019, and is now between 45-50% as the current wells have gotten younger.

At the current production rate of around 6.8mbpd, that means that around 3mbpd of capacity is lost each year and must be replaced by new wells for any future growth to be possible.

So the rigs which are working the Permian are mostly replacing existing capacity rather than expanding capacity.

However, rig count fell by a third from 2023 while production increased by 16% since then.

Why are decreasing rigs leading to increased production?

There were two contributing offsets which allowed production to keep increasing while rigs have been in decline since early 2023, both of which are rolling over.

The first is: longer laterals have been masking lower quality rock.

Shale wells run vertically down to the target formation (where the oil is), then rotate and run horizontally through it (the lateral). You can roughly scale up output to the length of the lateral.

So while you might see headline numbers being reported as output per well, output per lateral foot measures how productive your rock is.

Until 2021, production per foot and lateral length both increased, as the 2015-2016 and 2020 price busts forced everyone to operate their absolute best rock to remain in business. Goehring & Rozencwajg's work explored this idea in detail: a large part of the shale miracle was simply selecting higher grade rock.

Of course, superior engineering also played a role, with each foot getting cracked much harder, with more and more sand (to let the oil come through the rock, you frack it, which means you put high pressure water and sand until it cracks, and the sand keeps those cracks open for the oil to flow). Better technology and tools also enabled superior extraction.

But since 2021, that number has been declining, as the best rock has been declining. The Permian is 60% through its tier 1 rock, and every year the average new well is on worse acreage than the previous.

This has been masked by laterals getting longer and longer, from about 5,000 feet ten years ago to over 10,000 feet today. 15,000 foot wells are now somewhat routine, but these are close to both the mechanical and economic limit, meaning that they are unlikely to continue increasing forever.

The second offset was the completion backlog.

Shale wells are built in two steps. First the rig drills the hole, and then a seperate crew fracks and completes it later, at which point production starts. If you don’t complete it it’s considered drilled but uncompleted (DUC) and just sits in inventory. You can drawdown your inventory until you can’t.

In the Permian, DUCs peaked in 2020 as 3,500, and have been in freefall ever since, now down 75% to below 800.

(The chart below only shows since 2022 due to a change of methodology in counting at the EIA).

The DUC’s have mostly flatlined in 2026, as we are likely reaching the trough in the inventory drawdown. There are about 450 wells drilled per month, so you always have somewhere above 600 which are just waiting to be completed. Add in an extra buffer, and some dead stock, and the fact we went from drawing down DUCs at a rate of 49 wells per month in 2024, 43 per month in 2025, and only 11 per month in 2026, and the picture is clear: the drawdown is over.

So we managed to reduce the number of active rigs by always increasing the number of feet per lateral, and by using up the inventory of wells that were already drilled.

Both of those have expired now.

So what does it take to add back rigs?

Oil went super high earlier this year but we didn’t start adding rigs until May. Why? Nobody is drilling a well if they do not believe that WTI will be high enough to cover operating costs.

And like I said above, the quality of the rock has declined, so the breakeven on a new well has gone from around $50 in 2016 to $67 today. The cheapest inventory has been drilled.

And thanks in part to jawboning and the promise that “peace was just around the corner”, nobody believed that WTI would sustain a very high price through the year end amid the conflict. Expectations reset higher, but not dramatically so.

In 2022 when everybody expected oil to be well above the breakeven, more wells were drilled which brought in more active wells. In 2025, new-wells were unprofitable on average so less were drilled and active rigs took a nosedive.

In 2026, with the expectation now above breakevens, we saw the first quarter with a meaningful increase in average oil rigs since 2022.

So to wrap up everything I’ve covered about the Permian so far:

• Replacing output declines forces drilling or production cuts.

• Replacement means that even in Q4 2025 we barely cut rigs despite unprofitable margins.

• Reductions in rock quality and limits to efficiency mean more wells are needed to replace supply.

• We have drained the inventories of DUCs.

• The margin has now recovered and is likely to remain at a level which supports real activity.

But rigs are not all made equal

The top line analysis here has measured rigs as one fungible asset class. But treating the total count as a single market is a mistake which investors unwilling to go the extra analytical mile are leaving for us to pick up.

Drilling a 3 mile lateral through hard rock requires a specific class o equipment: enough horsepower and an AC electric drive to turn and steer miles of pipe, a walking system that moves the rig between wells on a pad without disassembly, and racking for the pipe itself.

These are known as super-spec rigs, and rigs below this spec are getting less demand. Of the 1,059 rigs in the US that are available, only 613 worked at all last year, and roughly 450 of these are super-spec.

When activity falls, operators start by releasing the lowest capability rigs first and keep the best ones working.

If activity picks up, then the demand for super-spec rigs is where all of the demand concentrates as the requirement for longer laterals can only be fulfilled by these.

Helmerich & Payne ($HP) is the market leader with about a third of these. But their operations span the full Lower US, Middle East, Latin America, and Australia.

So this thesis would be best expressed by a drilling contractor which owns only super-spec rigs, and operates primarily in the Permian.

There happens to be a stock which ticks all those boxes, pays you a 10% yield to wait, carries no debt, and is cheap as chips.

Of course this doesn’t come without a few risks.

It isn’t listed in the US, the dividend payout was cut by 72% in the previous downcycle, the controlling family has bankrupted a drilling company before.

Nonetheless, I think you will agree that this is an opportunity too good to skip.

How this idea came to be

Investment ideas never surface in a vacuum, and this one is no exception.

So before I unpack it, let me hat tip the gentlemen who were involved in me surfacing the idea.

I was talking on an X Space hosted by George Noble last month. During the call I was introduced to Matt Polyak of Hummingbird capital, who said he was particularly interested in oilfield equipment stocks.

The details of the thesis weren’t fully exposed, but it got me thinking:

If the Permian is to be the swing supply once we have emptied the SPR, then rigs come back, and this comes with orders for pumps, rig components, wellheads and other bits.

I didn’t act on the information, until I came across a post on X by Calvin Froedge, naming this little stock.

A lightbulb went off: Don’t trade the second order effect, trade the mispriced first order effect. Why go for the trickshot when you’re offered a couple free throws?

So thank you gents.

Introducing: NorAm Drilling

The company is NorAm Drilling, ($NORAM on Euronext Growth Oslo).

This is a Norwegian holding company, within which is nested a Texas Operator headquartered in Houston, which runs 11 rigs in the Permian.

The current market value of this company is around $170 million.

All 11 of its rigs are what the company calls “ultra super-spec” capable of drilling for both oil and gas.

Their fleet ranks near the top of the Permian’s rankings on feet drilled per day, and 7 of 11 rigs are contracted to major E&Ps.

Oh, and one rig even won rig of the year within a supermajor’s program… So they have the MVP of rigs.

The stock is as pure an expression of the thesis presented in this article you could hope for: No external debt, a two person executive team, just over 300 employees, and all-in cash cost of roughly $24K per day, making it the lowest breakeven in the industry.

They distribute all their free cashflow as dividends monthly. The stock currently trades at 42 NOK, and they currently pay out 0.45 NOK in dividends per month which annualizes to 12.9%.

The economics on the stock are a function of two things: utilization rate and the day rate on their rigs. Utlization is currently at 100%, which works out to about 3,930 rig-days per year. This means that each $1,000 of realized day rate add about $3.9 million of annual EBITDA against a $178mn enterprise value.

The company stopped reporting its day rates in June 2024, as it was the only company doing so monthly, and likely giving a clear signal to competition in a declining market of by how much they needed to undercut NorAm.

What is interesting though from the chart above though, is that we can see that the peak in day rates in the previous cycle came a full 10 months ater WTI peaked. Running a correlation against WTI confirms this: about 0% correlation against current price, but 0.8 against the price 10 months back.

NorAm’s utilization rate corroborates the replacement drilling thesis I established earlier. It continued drilling in 2025 at around 80% utilization whilst WTI was between $60 and $65. Premium rigs were hired at the bottom of the cycle because the drilling simply could not be deferred.

Utilization is at 100% or close to it, with the backlog a solid 16% of enteprise value. If what I said about realized day rates coming through with a log holds, then we’ve got a natural rise in EBITA which will flow through to shareholders as dividends over the next year.

The controlling shareholder

Geveran Trading and SFL Corporation own a majority of NorAm shares. Both are controlled by Fredriksen, who has a record of buying cyclical assets at a trough and taking investors along for the ride with him.

John Fredriksen was born in 1944 in Oslo, who left school in his teens and started as an errand boy at a shipbroker. He learned the oil trade in the 60s chartering tankers out of Beirut, and bought his first ships in the mid 70s slump.

He made his fortune in the 80s hauling crude out of Iran’s Kharg Island through the middle of the Iran-Iraq war, at great risk.

He built Frontline into a very successful tanker company, consolidated Mowi into the world’s largest salmon producer. His approach has been consistent over 50 years: buy cyclical assets near the bottom, distribute cashflows throughout the recovery, and take ownership of the same security class as minor investors.

His track record does have one chink in its armor, which comes in the form of Seadrill.

Fredriksen founded Seadrill in 2005 and built it, with debt, into the largest offshore driller in the world, worth more than $20 billion at the 2013 peak and carrying about $14 billion of liabilities.

When oil fell in 2014, the debt forced Chapter 11 in 2017; shareholders kept 2 percent. The recovery never came, because shale had taken offshore's role as marginal supply, and the company filed for bankruptcy again in 2021. Shareholders were wiped a second time, and this time Fredriksen let it go.

Mistakes force lessons, and I believe, with NorAm, these lessons were learned: financial leverage on-top of operating leverage is what will kill a drilling company, a drilling cycle can be misjudged for years, and you should not expect the owners to defend a losing position.

How is this thing so cheap?

At 42 NOK, as I mentioned above EV is around $178mn. I always reverse engineer valuation rather than project forward.

I try and answer the question: Given the current price, what growth rates and conditions is the market pricing? If my thesis disagrees with the market pricing, we can comfortably enter the position.

Let’s assume an 11% required return, which is appropriate for an illiquid, cyclical, microcap. That would assume $22mn in EBITDA. 2025, which was the worst year on record generated $20mn in EBITDA.

So the market is currently pricing the worst year on record as the permanent state of the market.

If we witness anything like a return to 2023 conditions, when NorAm generated $40mn in EBITDA, this stock would be worth double what it is today.

Of course with cyclicals it is always difficult to pinpoint valuations, but NorAm is trading at 8x trough EBITDA.

The stock repriced to NOK 53 in May before declining on the back of a perceived piece in Iran.

From the technical front, 2026 has marked a clear move away from the descending channel of 2023-2025, and while it’s still early, we seem to be seeing an emerging trend line which should give us support in the $38-$42 range throughout the third quarter.

I’d place a stop loss at NOK 35, which would be a break from the current environment and a return to the prewar pricing. If price goes there, we likely want to be out and reconsider what is happening. 17% below the current price means you can size at 5-6% if you want to put no more than 1% of your capital at risk.

I believe the market is pricing extreme bearishness on the conditions in the Permian which simply do not add up with a $70+ oil world.

At its 2023 peak, NorAm paid $4.7mn per month in dividends, but by late 2025, it was only paying $1.3mn. If WTI settles in the low $50s, then we’re looking at going back to those days and to the lower valuations.

I just don’t see that happening. I think that the current conflict, which let’s not forget, is the largest supply disruption in the history of oil, likely puts a floor under oil prices for a while. That’s what the price action has underwritten so far.

How long will China underwrite oil prices for the entire world by drawing down on its massive SPR?

Wrapping it up

I believe this stock is a perfectly fine way to play the world we are entering in. You get a double digit yield, optionality on a rerating to give capital returns of 50% to 100%, and a downside which is capped if you believe that the US Iran woes are more complicated than previously marketed.

If you enjoyed this article and it was of value to you, I will ask a couple things of you.

First please make sure if you haven’t yet to subscribe to this newsletter. You can opt-in for free and continue to receive all the free content I will be sending your way weekly for the next month.

Subscribe now

You can also decide to pledge a subscription.

I am still deciding on the ultimate prices for the subscriptions.

In the meantime I have left the pledges well below whatever the launch price will be, as I want to do a little something for early readers and people who get with the program early.

So if this was valuable to you, subscribe, pledge, and share to someone who will enjoy this article.

As Babylon Burns, we’ll light the cigars.

Share

Actually, single asset, or sector permabulls are mostly idiots.

Investing works in seasons. Certain assets do well in certain environments, and others do better in other environments.

Being attached to a single asset like it’s a sports team, is a weird, puerile,

In this article, I will show you exactly why a lot of “investors” are misguided in their unabated bullishness in gold.

I’ll show you what, how, and when to make the most money in the gold trade.

Sound good? Let’s dig in.

(PS: before we do, if you haven’t yet, consider subscribing for free to the newsletter. Until August 15th I will be giving away a weekly supply of FREE in depth articles exposing my thesis, how I think about different investment opportunities, going from the macro read to the micro minutiae.)

When J.P Morgan testified in front of Congress he made a very clear argument. Money is gold, and nothing else. Money is a scarce, fixed, physical, finite thing, whereas credit is fluid, judgment based and relationship driven.

He was arguing that credit wasn’t something which was mechanical and controllable, and therefore wasn’t money.

Fast forward 114 years later, and we are now 55 years into a full-fledge fiat monetary system, where most of what we call “money” today, is nothing more than a bank deposit, or in other words, credit which depositors extend to their banks.

What’s more is that since 2009, we have moved from the private sector being the multi-node controller of credit creation, to the Fed having the most discretion over how much money moves in and out of the system.

Consider that in the US, since 1959 while real GDP has 7xed, broad money (M2) has 75xed.

When we see the sharp changes in the amount of money in supply which occurred after the Great Financial Crisis of 2009, and then again in response to 2020 lockdowns, it is fair to say that those who control the money have not been doing so with your best interests at heart.

Money that is printed in excess of what the economy demands needs to go somewhere. It can go into prices (inflation), it can go into assets (stocks, gold, houses, land, bonds), or it can sit idle.

Gold is a brilliant asset to hold

And governments debasing money is a tale as old as time. If we price gold in sterling pounds going back to 1257, we can see that that over the course of the first 500 years, gold 5xed against the pound.

Between 1971 and today, it 180xed.

I explained in my first article on Substack why governments worldwide would have no choice but to continue doing that, and why scarce assets, like gold, were imperative to hold.

Like Ray Dalio said: If you don't own gold, you know neither history nor economics.

And I do believe there is a certain amount of gold you want to hold yourself. One of the biggest draws of gold is that it has been used as currency for a very, very, long time, and that it does not depend on a counterparty.

With fiat, you depend on the corporate at which your deposits sit, and the sovereign which underwrites and dilutes the currency.

The gold bug in me believes it’s a good idea to own some physical gold: coins, jewellery, small bars, which you self-custody. This way, you depend on no other counterparty, gold you wear can move through borders more easily, and you have part of your wealth stored as a hedge against the current status quo.

But gold miners are not gold.

When I present this long term thesis and view to my peers, I am often met with the following question:

So what’s your favorite gold miner?

And they are surprised to hear that right now, my exposure to gold miners is absolutely zero.

Gold miners are essentially a leveraged bet on gold which as you’ll see in this article decay when gold is either down or sideways.

A miner sells ounces of gold at market rates and spends a largely fixed amount to produce them. The spread is their margin, which bakes in considerable operational leverage, as a smaller move in gold leads to a disproportionate move in margins.

It’s no surprise then that miner ETFs have gold-betas of around 1.7x with the higher volatility names having betas of more than 2x.

In gold uptrends, miners have done better than gold, and in downtrends they tend to do worse than gold.

So you are dealing with instruments which amplify gold’s returns.

So why has holding them been a losing bet?

When you hold a gold miner, you hold a claim on its margin.

From the 2011 peak, gold price went from $1,900 to $4,100 or so today. If we look at Newmont ($NEM)’s all in sustaining cost (AISC), it went from $1,100 to $1,600. (although the concept of the AISC was first defined in 2013, the numbers still pretty much check out)

The margin went from $800 per ounce to over $2,575 per ounce. You’d think the margin exploding would imply that the stock would have returned much more than the return on gold?

Unfortunately, that would be wrong. Newmont returned 110% while gold returned 130% since the 2011 peak.

Most miners in fact, did worse than the metal, while the royalty companies did better.

Unfortunately the AISC masked things that could have gone wrong for investors and which ultimately did.

First, mining reserves deplete every year, maybe to the tune of 5-8% on average, and these need to be replenished with new assets. Growth CapEx does not make its way into AISC. In 2025, Newmont deployed $3.3bn in CapEx, but only $1.8bn got classified as sustaining. The boundary between what is sustaining and growth is blurred enough that even brownfield expansion of an existing mine can be classified either way.

So the industry standard reporting metric isn’t perfect and understates the margin.

The bigger problem is this one: Newmont’s gold production went from 5.2 million to 5.9 million in 2025, which was a 13% production increase in 14 years. Over the same time frame, share count went from 494mn to 1,068mn, so production per share actually declined by 48%.

Production is expected to decline to 5.3mn in 2026. Now to be fair to management, they are now buying back shares, and even adjusted for the hidden increase in costs and the dilution, based on prior peaks we could have expected more of a rerating than we did.

But by and far, over time, gold miners eat themselves, replacement costs more each cycle, this replacement bill gets paid with investors money, the operational leverage is diluted by increasing costs, and you’re exposed to human bias which torches cash.

Does that mean that I’d never buy gold miners?

No, and I want to be clear on this.

Gold miners are a structural losing proposition over time vs holding gold. So for passive long term investment, you want to prefer gold every single time.

But when you are riding a trend in gold, then trading miners can make sense for the amplified returns you can get versus the metal.

So when I do decide to buy gold miners for the trade, I am not so interested in buying the senior high quality ones. Buying Agnico ($AEM) gives you low-jurisdiction risk, reliable guidance, cost discipline, and defensive downside behavior relative to the sector.

All of those things are great if your plan is to hold throughout the cycle. As we have seen, there is little justification for holding gold miners throughout the cycle.

As the chart above shows, in such setups, it is the higher risk names like B2Gold, Harmony, and Iamgold which have cheap optionality embeded due to jurisdiction risk, turnarounds happening, or higher operating costs.

Beta to gold is one thing, correlation to it is the next you’ll look up. If you’re going to trade an explosive up move in gold, you want your levered miner to be as pure an expression of gold as possible. The upper quadrant of the chart above gives a decent universe to work from.

And I know someone is going to say “but gold miners are exceptionally cheap, and printing cash at the current rates.”

Yes that’s true. But as a trade, they are ultimately just gold beta, and in an environment where gold remains sluggish or trades sideways, they might be dead cheap, they’ll also be dead money.

What would it take to go long these miners?

For me to get excited about investing in gold miners, I would need to see movement in gold.

How do you value gold? Difficultly.

In Berkshire’s 2011 annual letter, Warren Buffett tried to do so with the following thought experiment.

He was obviously vindicated as the S&P 500 did a lot better than gold did since then.

Gold is not productive. It is a function of fear, and a function of a reduction in value of the thing it’s measured in: dollars.

And gold works in spurts. It does nothing for a while, then it has a massive leg up..

And recently, it reached a peak, rolled over, and dipped below its 200 day moving average.

In an interview in 2014, Paul Tudor Jones once said:

“My metric for everything I look at is the 200-day moving average of closing prices. I’ve seen too many things go to zero, stocks and commodities. The whole trick in investing is: “How do I keep from losing everything?” If you use the 200-day moving average rule, then you get out.”

The 200 day moving average isn’t a perfect line in the sand, but it can be comfortably used as a gauge of: is something trending or not?

And the answer here is no.

Technical analysis is beautiful for assets which can be difficultly valued because the chart gives us a pictorial record of the tug of war between buyers and sellers.

As Richard Schabacker, the modern founder of technical analysis said:

In the record of such trading all of these many and varied fundamental factors are brought to bear, are evaluated and automatically weighted and recorded in net balance on the stock chart. The trading in any stock is largely the result of the influence these fundamental factors have had on each buyer and seller of each share of stock. The stock chart is a pictorial record of such trading, so that it, in itself, is a reflection of all those other factors and, from a purely technical standpoint, need therefore concern itself no further with such fundamental considerations

I like to look at formations which build up over time. That’s where technicals prove value to me, as they give me a good idea of the general sentiment towards an asset.

And Gold was basing in a rectangle between 2021 and 2024, mostly between $1,700 and $2,050. Then in 2024, it broke out and commenced a bullish ascending channel which brought it all the way to a blow of top at $5,318 on January 30th this year, just as Kevin Warsh was nominated chairman of the Fed.

Since then, it broke out of that channel (around the same time it broke below its 200 day moving average, which is interesting confluence), and has been forming a down-turned wedge ever since.

Here’s the good news: Down turn wedges are reversal patterns, so once they resolve, we expect the previous trend (up) to resume.

Here’s the bad news: They usually resolve very slowly, often staying within the formation until the apex is reached. And once a breakout happens, they tend to move horizontally for a while before a clear trend resumes.

You can view this as a period of lull when buyers and sellers are uncertain of quite where the asset is headed after a strong impulse up, and an aggressive correction down.

On the topic of this formation, Schabacker said:

“It presages a reversal of trend, but the move is delayed for an indefinite period in the case of many Down-turned Wedges. The practical trader will therefore take prompt action when his chart shows a breakout from an Up-Pointed Wedge Top; whereas he may safely delay action at a Bottom until the price and volume action indicate that the new uptrend is really started."

This formation could continue to play out until the fourth quarter this year, and the climb back from there could be slow for a long period of time.

Given that we have assessed that gold miners do well only during gold up regimes, we can comfortably conclude that for the time being, we don’t want to touch them with a 10 foot pole.

To be clear, technicals are not forecasts, nor do they inform us of probabilities, they only give us possibilities. And the possibility which is being painted by the current record is one where it will take a while before gold makes new highs.

Will I let you know when I change my mind?

Yes of course, that is why I’m launching this newsletter. On August 15th, on the 55th anniversary of Nixon moving away from the Gold standard, I will launch the paid version of Babylon Burns.

In the meantime until then I will be giving you a weekly drip of (I believe) high quality articles 100% free. You just need to put your email in below if you haven’t already.

Subscribe now

Once I launch the paid version, for annual subscribers, it will include a web and mobile application with the portfolio, the watch list, and a custom chart library with my takes on assets, stocks, and so on, and the formations I am waiting to enter into positions.

I am getting clearer on what the offering will be although this isn’t set in stone yet.

• The monthly tier will be priced higher, as I don’t think these newsletters are made to be followed monthly.

• The annual tier will be priced between $299 and $499 and include the app mentioned above.

• The Pro tier will be priced between $1,000 and $2,000, will be limited in spots, and include a quarterly call with me, a signed copy of my upcoming book and other perks.

For the time being, I have left the pledges well below those levels, as I want to do a little something for early readers and people who get with the program early.

So if you decide to pledge to a subscription now at the lower rates, I’ll honor those for you when we go live, even if the official rate is way higher.

Subscribe now

In the meantime, please share this with anyone you think might benefit from reading it.

As Babylon burns, we’ll light the cigars.

The next day, Zuck made very clear the rationale behind this decision when he said:

“The trajectory of the agentic development over at least the last four months hasn't really accelerated in the way that we expected”

The stock popped from $560 to $628, before giving back those gains and closing at $582 on Thursday.

And I think that price action on Thursday was telling and that the cracks are starting to show in the entire complex.

Now to be clear, I have been an AI bull earlier than most.

Among other picks, I’d been long Broadcom ($AVGO) since 2020 at $35, and made it the largest position of my portfolio in 2023 adding shares at $60, when I wrote a piece on SeekingAlpha saying that “ChatGPT marks a turning point in the AI revolution”.

I still have a position on what has been an 8 bagger. I am not a Jeremy Grantham or any other permabear who is salty because he got it wrong and missed the bus.

Subscribe now

Yet, I’m becoming very concerned.

Why bother with timing at all?

As everyone knows, the difficulty lies in the fact that so much money stands to be made in these irrational bull markets, that it can be difficult to step back and say “enough is enough”.

Everyone knows we’re on a bus heading for a cliff, but if you get off the bus too soon, you’re just as wrong, if not more, than the guy who hurled himself off the cliff.

Like the chart below, (which I shared on X the other day) shows, the 2002 trough for the S&P 500 was 39% higher than when Jeremy Grantham called stocks “dramatically overpriced” in 1995.

There are two quotes that come to mind.

The first, from Peter Lynch:

Far more money has been lost by investors preparing for corrections, or trying to anticipate corrections, than has been lost in corrections themselves.

And the second from Howard Marks:

Being too far ahead of your time is indistinguishable from being wrong.

So one might ask why not just stay the course, and stick around for the long run?

Well, there are a few reasons. First is, in the long run we’re all dead. Second is, you have to do some soul searching and ask yourself if you truly have the constitution to endure 50-70% multi year drawdowns in your portfolio. Everyone has a plan, until they get punched in the face, as Mike Tyson liked to remind us.

Third, avoiding one nasty drawdown will do so much good for your net worth. (You can look up “On Minimizing Drawdown risks of lifetime investments” and “large drawdowns and long-term asset management” for the two papers on the topic).

Fourth, Peter Lynch’s comments came at the tail of a career which spanned 1977 to 1990, during which the nastiest correction happened between August 1987 and December 1990, when the S&P 500 corrected 33%. The 2002, 2009, and 2020 drawdowns all eclipsed the 1987 one.

We’re living in a regime of heightened volatility, of large governmental fiscal debts, of a massive yen carry trade providing marginal liquidity, of multipolar fracture. Everything suggests that the path forward is a volatile one.

There is real money in avoiding the unwinding of the AI trade.

Grab yourself a cup of coffee, and let’s get started.

When it all falls down.

The current AI buildout is based on a certain set of assumptions which impact the stocks at various parts of the value chain.

And a lot of it rests on The Jevons paradox, the nineteenth century observation that making coal engines more efficient increased coal consumption, because cheaper power found endless new uses. This has fuelled a large part of the AI buildout boom.

The baseline assumption is that every improvement in models, every decrease in the cost of tokens would expand demand as more tasks and workflows become economically viable.

Based on that logic, the appetite for computing is bottomless and unlimited capital expenditure is rational.

Based off of this assumption, the hyperscalers (Amazon, Google, Microsoft, Meta, and Oracle) have guided roughly $690bn in capital expenditure in 2026, up 81% YoY and 207% higher than in 2024. They’ve committed more, and will effectively be turning free-cashflow negative very soon.

If the appetite for compute is infinite, then this makes sense.

But the Jevons paradox theory only makes sense if the cheaper inputs produce something which is actually worth buying. So underlying the whole spend is a massive assumption which is largely untested at this point: spending money on AI generates positive returns.

And unfortunately, this underlying assumption is starting to crack, and you’re hearing this from somebody who believes that the AI innovations we’ve seen since the launch of ChatGPT are most likely one of the most revolutionary technologies EVER.

I’m sure at this point, we’ve all read the MIT study (or at least the headlines) of enterprise AI pilots, that found that 95% of generative AI pilots delivered zero measurable PnL impact, against $40 billion of enterprise investment.

There are some fair critics of the methodology, but the direction is clear and validates my premise that I explained on George Noble’s podcast last week: AI induced productivity, and therefore AI returns, will take years, if not a decade to arrive.

This is well grounded in economic theory which I will only gloss over here.

Brynjolfsson, Rock, & Syverson wrote a paper which posited that new general purpose technology will follow a productivity J Curve, which is to say that productivity goes down while firms redesign processes, retrain staff, reshuffle organizational structures and data pipelines.

The AI pilots are most of the time ad-hoc investments made by corporate America, the vast majority focused on sales and marketing, which concentrates probably the lowest leverage implementation of the technology.

Carlota Perez has a similar framework of installation vs deployment phases of a new technology, with the added twist that there is usually a financial bubble then crash that happens during the installation frenzy. Sound familiar?

OpenAI and Anthropic are burning through cash as they engage in an arms race. Profitability is at best years out.

The whole argument of measurable returns that come in time is really coming into question when Meta says it has “excess compute”. Now Meta is a firm which arguably should have infinite hunger for compute, with 3.5 billion users, an advertising machine that converts intelligence into revenue, a lab of AI researchers which were hired at any price, and 15,000 lay-offs which were meant to allow for the massive investment into AI.

Back in May Zuck said that selling compute was definitely on the table if Meta overbuilt capacity, and just two months later is saying that “agent development did not accelerate as expected”.

The bulls found a way to spin this. Jefferies’ Brent Thill brought up the Amazon comparison: Meta is following the Amazon Web Services playbook that turned spare servers into the most profitable infrastructure business EVER.

I don’t buy it. Amazon created a market that didn’t exist and grew it for 20 years, Meta now is adding supply into a rental market which is already seeing questionable profitability. That was a euphemism.

The GPU hour will be the first casualty

By launching a cloud business to sell excess compute, Zuck is admitting that Meta overbuilt.

Selling compute was the Neo-Clouds business model. Nebius ($NBIS) and CoreWeave ($CRWV) financed the buildout of servers with debt that cost 9% to 15%. But the price of an hour of GPU is down from $8 in 2024 to $2.99.

On their books, they say they will depreciate over 6 years, which makes the buildouts profitable above $2.24 per hour. The bear case has been argued for the longest by Michael Burry, who says that GPUs have an effective life of 2 to 3 years, which would make them unprofitable in any scenario.

But the market price of GPU hours is a commodity. It is a function of supply and demand, and Meta which was one of the biggest sources of demand (it has a $35bn commitment with CoreWeave) is now becoming a source of supply.

What does that tell you about where the price is going?

Beyond its $35bn CoreWeave commitment, Meta also committed $27bn at Nebius. When the customer becomes a reseller, no matter what you want to believe, you know that those contracts were the ceiling on these companies. Sure the CoreWeave contract is take or pay until 2032, but under what economics? In general CoreWeave have been discounting multi-year contracts, and have used these contracts as the collateral for the leverage they have contracted.

If the price of H100 GPUs flattens or declines with marginal demand turned marginal supply whilst the useful life of the chips is below their expectations, these guys are toast.

Nvidia is the next (circular) link.

The continuum of the already well made by others argument goes that once Hyperscaler demand flattens, Neo-Clouds buckle, then so does Nvidia, because it was not only the vendor of these chips, it was their financier.

Nvidia had committed up to $100bn to OpenAI, which it then cut 70% to $30bn just before Jensen Huang went on stage and declared it and a $10bn Anthropic stake the final direct investments made by the company.

Beyond OpenAI and Anthropic, Nvidia has committed $20bn to xAI through an SPV, has arranged a $6.3bn backstop for CoreWeave stating it will rent whatever capacity they cannot fill, plus they’ve taken $2bn in equity in each of Coreweave, Nebius and IREN, plus a $1.5bn sale leaseback with Lambda in which Nvidia became its own customer’s largest customer.

Vendor financed demand is literally the lowest quality demand that exists, and they stopped doing it in March, just a quarter before one of the marquee buyers announces it has spare capacity to sell.

NVIDIA has written insurance on the utilization of its own product, which means a falling rental rate is no longer only its customers’ problem.

While I usually rebuild all chart myself with my branding, I had no interest in rebuilding the circular economy chart which would have been a nightmare.

Thank you Bloomberg for the chart.

The Memory cartel gets hit next

The unwinding runs upstream to the most cyclical layer in the chain: HBM/DRAM.

The stacks of memory required alongside GPUs are built by an oligopoly: Micron ($MU), SK Hynix, and Samsung.

In a frenzied buildout, the supply has not had time to adjust, with spend on memory going from 7% of data center costs in 2023 to 30% today.

I came across an interesting chart from Andreas Steno Larson the other day which showed YoY growth in Korean exports as a decent coincident indicator for semiconductor stock prices. I built it adding on the forecast of Korean semiconductor exports going forward.

These charts are cute until they decouple (remember the global M2 and Bitcoin chart?), and the decoupling usually comes to the downside when the stocks turn more negative than the exports.

Nonetheless they do bring up an interesting discussion point, which is that assuming fixed prices, given that we know the supply is pretty much fixed until late 2027, then the arithmetic suggests that the growth rate in Korean exports must peak between now and October. After that the arithmetic just doesn’t support it.

The price curve after having undergone some dramatic changes, is now showing signs of flattening for DRAM, HBM, and NAND (to a lesser extent)

I will not pretend to be able to pinpoint a peak in contracted pricing, but matching a growth rate is a difficult thing to do.

As we roll over into higher comparables, it gets harder to put up crazy YoY growth numbers. As the growth numbers come down, so do the expectations, so do the valuations.

And in 2027 and 2028, new fabs are coming online, which tells us there is a light at the end of the shortage tunnel.

Unfortunately for the stock prices, that light might be freight train hurling towards them.

The hyperscalers come last

Unlike the neo-clouds, the hyperscalers do not face a solvency question, since they’ve been funding the buildout from some of the largest profit streams in the history of capitalism.

The problem is if the returns don’t live up to the demand, then they will be left with investments which will not produce the returns that are expected over assets which depreciate probably quite quickly.

What we’re going to look for are the first indications that CapEx is flattening or declining, but this will of course be a lagging indicator.

If the returns don’t come from the CapEx, these firms will cut the CapEx to protect their cashflows, after having turned them negative in the build-up.

This will necessarily come with an unwinding of prices, but it might be less extreme in scale than in the layers above.

Reading the tea leaves

This brings our analysis of the future chain of events to an end, which brings me to the usual conclusion:

“Cool Story Sam, but what do the charts say?”

I believe the validity of classic technical charting is now well established. I tend to keep my analysis simple, abiding to the principles which Schabacker set out in the 1930s.

Back then he posited that charts show us “the balance between buyers and sellers, the balance between supply and demand for any stock or group of stocks; in short the phenomena arising out of actual trading, per se, in a free and open market."

I agree. Price is a log of how buyers and sellers felt about all known factors. In price we see exuberance, trends, exhaustion and reversals.

I am not a pure chartist, my approach is one of confluence. Do the numbers, the narrative, and the public log (price chart) of the events converge in the same direction?

With the necessary lip-service to technicals out of the way, let’s proceed. I will be looking at charts at various time horizons and asking: what do they tell us collectively?

The Neo-clouds, whose business is to finance compute with debt and sell it to the hyperscalers like Meta, saw horrendous price action on Wednesday and Thursday.

Nebius Group ($NBIS) fell from $258 to $215, falling out of its 2026 ascending channel in the process, flipping the trend negative.

IREN ($IREN), the converted miner, completed a head and shoulders pattern which confirmed the reversal on June 24th. It continued its descent from $45 a couple days ago to $38. The default target from his is a retest of $27.

Coreweave ($CRWV) which in April had staged a breakout following 3 quarters of consolidation, broke back down during the two trading sessions.

It would be natural for it to head lower, the trend has also shifted bearish in the past two days.

Nvidia ($NVDA) which financed many of these NeoClouds, is sitting on the lower boundary of the massive rising wedge that has been forming since its historical run from $38 in 2024 to $234 in May this year.

Yogi Berra reminds us that it ain’t over ‘til its over, but this bearish formation has an end date on it: Q1 2027. The chart is telling us a simple story: either NVDA has topped and it will fall below $190 soon and confirm this, or it has another leg up, that could be as high as $250-$275, that will happen within the next 6 to 9 months, and then resolve lower.

Micron ($MU) despite one of the most upbeat earnings calls ever is seeing its chart break down too. The historical run from the $310 level in April to $1250 formed a bearish rising wedge, not unlike the one in NVDA, but on a shorter time frame.

In the past two sessions it has broken down clearly from the support. In June it tried but failed a breakout to the upside, and now everything is reversing.

The KOSPI, the Korean Stock Exchange, of which two chip stocks are more than half of the index, has broken below the ascending channel of the very same move that powered Micron since April this year.

$7,375 is the line in the sand. Below that, this trade unwinds, and along with it, the leverage which Korean retail has taken on to participate in this trade.

Move to the hyperscalers, Microsoft ($MSFT) broke down of the wedge it has been forming since 2022, retested and got rejected.

The $350 level marks the line in the sand, below that the stock turns very bearish. At the best it’s looking like sideways consolidation between $350 and $450, the downside is particularly bearish.

Oracle has been forming a consolidation triangle, and currently it sits at the very bottom of it, slightly below.

Now technical charts only inform us of possibilities not probabilities, but because there is no trading on Friday, this would mark the first weekly and daily close outside of the formation, suggesting further bearish sentiment from here onwards.

Google ($GOOGL) is more nuanced. The good news is that it seems to have broken out of the epic ascending channel that started in 2009 at an adjusted price of $6 bringing it all the way to $359 today.

The bad news is that it is staging such a breakout by forming a clear ascending wedge, the same bearish formation which Micron and Microsoft broke down from and which is threatening Nvidia. This means the breakout is likely to be a blow off top which reverses hard.

Unfortunately the bad news does not end there. Amazon ($AMZN) is also approaching the APEX of a rising wedge within the next year.

Schabacker said of rising wedges that they represent “a shrinkage of activity in the stock as the feeling grows that it is temporarily over-running its true relation to underlying values.”

I fear, after running through this chart library, that this is exactly what we are seeing with the AI trade.

Nothing good happens after 2 AM.

I grew up watching the TV show How I Met Your Mother. My brother and I would binge watch it on Megaupload (the good ole’ days).

In it there’s an episode where the main character, Ted, tells you there’s nothing good that happens after 2AM. The party might rage on beyond then, but people are more irrational, mistakes happen more easily, and anyway you look at it the risk return just doesn’t look right.

That’s where I feel we are with the AI trade.

It’s past 2AM. No matter how you look at it, the cracks are starting to show. It makes sense that the bigger the new technology, the more it gets overbuilt.

This is the time to harvest gains. It most certainly isn’t time to deploy more capital into the trade. If you still have positions (as I do in Broadcom) that are at huge gains, it is fair to let them ride out until the tides have clearly turned.

I mapped out the key levels for most of the complex, and it should be clear this isn’t a “sell everything AI” call.

But a lot of the dominoes are starting to fall. I wouldn’t touch memory or neoclouds here, and would be very cautious about the rest of the value chain, and read the tealeaves carefully when choosing to scale out of positions.

What comes next

If you liked this article then make sure you subscribe as for the next 6 weeks I’ll be publishing reports absolutely free.

Subscribe now

Then on August 15th I will launch the paid version of Babylon Burns, which will include my global macro views, stock picks, and price targets in this coming regime. If you haven’t read the two previous pieces, I suggest you take a look at them at the following links:

• They sold you a dream and you bought it wholesale... (Introductory piece)

• There’s a 50%-margin monopoly hiding inside the airline industry that’ll print cash for the next ten years

Also I’m still deciding on the exact pricing of the newsletter. In the meantime I have left lowball pledges in place which I will honor for anyone who pledges before the newsletter goes live.

I’m not quite sure what price will be at launch, but there is no way you’re getting a better price than this.

As Babylon burns, we’ll light the cigars.

On August 15th 1971, Nixon abolished the gold standard and ushered in the era of fiat money.

This marked the beginning of money being printed in excess of growth in the real economy.

Whilst in 1971 a household’s median annual wages could buy 36% of the median American house, today that number is down to 21%.

The household’s wages could buy 89 units of the S&P 500 in 1971, only 11 today.

The Household could buy 258 ounces of gold, or pay for 20 years of public college tuition, vs 21 ounces and 7 years today.

To make it worse, a household today works 30% more than in 1970s. More women have joined the workforce, only for the household to be worse off still.

Now of course, this is good for those who own houses, stocks, gold, and operate universities. And not so good for those who save US dollars (or any of the other major currencies by the way).

It’s because when they print money in excess of what is needed to grow the economy, they reduce the value of money. It’s as simple as that.

But, there are things which cannot be printed. Things that are relatively scarce

Scarcity is the heart of economics. Price is set where supply and demand meet. If demand is scarce relative to supply, price craters, if supply is scarce relative to demand, price explodes.

It is why, in last week’s article within which I presented my framework, resource scarcity, is identified as one of the secular themes which will dictate the regime within which we are.

Resource scarcity evokes oil, lithium, gold. And those are all interesting, they are geological scarcity.

The discussion then often segues into Bitcoin, Swiss Francs, or any other discussion on sound money.

And that is fine. To be clear, I love the idea of Bitcoin, what it represents and its potential. But it is an asset which can only be traded only on sentiment. And sentiment, through price action, currently shows a textbook distribution, along with the price dipping below both its prior cycle all time high, and below its 200 week average price.

The risk reward isn’t anything I’m excited by here.

Which is the point I want to make: investors shouldn’t be married to an asset, but they should be married to the ideas which made that asset desirable. In this case it is scarcity.

And I think that investors should take the following frame:

That scarcity can be structural (Rolex or Hermes keeping supplies tight to manufacture high prices) or it can be tactical (DRAM prices exploding as demand moved faster than you can build fabs).

Identifying sustained scarcity which hasn’t been priced, is a great way to produce returns.

Today I will walk you through an industry which has structural demand and faces scarcity throughout the next decade: the aerospace after market.

The aerospace after market

“The engine is the heart of an airplane, but the pilot is its soul." - Walter Alexander Raleigh

Whilst the quote above is undeniably beautiful, the truth is that there is an oversupply of entry level pilots.

Commercial aircraft on the other hand, are not so plentiful.

My wife and I get to fly quite a bit. In any given year, we might spend between 7 and 10 days (200 to 240 hours) in airplanes.

Below is my 2026 H1 summary, which at the very least, should reassure you that you are reading global macro commentary from someone who has seen a fair bit of the globe.

Some flights go wonderfully well. Some, less so. Problems can come from two areas: poor service, or old, less comfortable planes.

We’ve noticed there are a lot more old planes than we would have liked. As is often the case, an offhand comment from my wife sent me down a rabbit hole.

It turns out it is because the world cannot build new aircraft fast enough, so it is forced to keep the old ones flying.

Planes which were supposed to be retired years ago are still in the air because there is nothing to replace it, and this will be true for most of the next decade.

And old airplanes (not unlike old humans) need more maintenance. And the heart of the airplane, as Mr Raleigh reminds us, is its engine.

Engines are a unique business model. They are generally sold to airlines around, or even below, cost. The manufacturers essentially give the engines away. This makes sense, because engines should be thought of more as multi-decade subscriptions.

During the 30 years that an engine remains on the aircraft, it must visit a shop every few years for an overhaul, which costs millions of dollars in parts and labor.

In most cases, the manufacturer controls the parts and is the only party legally permitted to certify maintenance work or supply the components.

The airline is captive, and has no choice but to pay (or have a grounded airline).

It’s a pure Gillette razor and blade model, where you’re willing to not make money upfront so that you can then cash in for a very long time.

But unlike Gillette, there is a regulatory component which constrains these operations to a small subset of players which dominate the market. Engine maintenance is not discretionary as it is mandated by law, which enables part manufacturers to pass through any increases in costs which they might incur along the supply chain.

The longer an aircraft stays in service, the more it needs maintenance, and the more onerous that maintenance becomes.

So the big question is simply: is the world’s fleet getting older and will it stay that way?

The answer to both of those questions is yes.

A shortage with no end in sight

The average commercial airplane today is about 15 years old, which is the oldest the global fleet has ever been. Before the pandemic, record deliveries brought the average age down during the 2010s. Since then deliveries collapsed, and with them the retirement rate of aircraft.

Global deliveries by Boeing and Airbus have undershot targets every single year since 2018, The two companies have been held back by their own supply chain constraints along with labor shortages (see scarcity causing the economic mismatch, here again).

The backlog of orders is now above 17,000 aircraft, which is the largest in history and will take a decade to clear. While production hurdles have been easing, engines in particular remain in short supply and will be for the next decade.

The backlog represents 58% of the fleet currently flying, versus historical averages of around 40%, which is a clean way of visualising that planes that otherwise would have been retired are still flying.

With a record old-fleet, retirements at record lows, production that is struggling to recover and backlog that cannot clear for a decade, it is no surprise that aircraft are being flown more hours than ever in their history in order to cover the shortfall.

The conclusion is quite simple: the installed base that requires maintenance is large and locked. This will not resolve until sometime in the 2030s, in the most optimistic scenarios.

The money flows to engines

The maintenance, repair and overhaul (MRO) market was worth about $136bn in 2025, and expected to approach $193bn within 10 years, which would be double its level in 2019.

This growth is being driven primarily by aircraft age, causing more maintenance per aircraft, as the fleet itself is growing at the slow, supply constrained rate I highlighted above.

Since 2019, spending on maintenance has increased by 40% while the fleet size itself has only increased by 10%.

Every airplane is becoming more expensive to keep flying, which is the exact dynamic we would expect from an aging fleet.

Within the MRO market, the money pools in one segment more than all others: engines.

Based on Oliver Wyman (a management consultancy)’s numbers, engine work is more than half of all maintenance spend, and is the fastest growing segment within the industry.

Of all repairs, engines are the most complex, the most tightly certified, the most mission critical, and as a consequence they are the most expensive part of an aircraft to maintain.

And here is the thing: the shops that are able to do this work are full, and there is simply not enough capacity to meet demand, and there won’t be for years. When supply and demand don’t meet, price moves to clear the market.

Bain & Co suggest that if maintenance capacity grows at its current pace, demand for engine shop visits will exceed supply by more than 17% by the end of the decade.

This is a market which cannot easily clear, because there is a massive regulatory barrier to surmount just for the right to participate.

You cannot maintain or repair parts for certified aircraft engines unless you hold the right certifications and approvals, or even the intellectual property that govern a specific engine.

So whilst the incumbents are struggling to ramp capacity fast enough to meet demand, there are massive regulatory hurdles to new entries.

This creates a long lasting setup for the industry, which will benefit investors who are able to enter these trades at the right time.

And I happen to have identified 3 stocks within the theme which make intuitive sense on the micro level, fit into the macro perfectly, and which have a chart which is looking constructive.

What’s more they are spread across 3 currencies (USD EUR GBP) providing some varied FX exposure. In all my writing, my baseline assumption will be that we are operating in US dollars. The vast majority of my income comes from the USD or currencies which are pegged to it, so it serves as my baseline. Everything will always be done to maximize USD value.

Trade 1: own the engine program

Safran (SAF) is a French aerospace company, listed in Paris, which through a 50-50 JV with General Electric , co-owns the most successful jet engine franchise in history: CFM.

CFM’s previous engine, the CFM56, has been built more than 33,000 times since 1982, making it the most produced jet engine ever. While production has ended, the installed base is going through some of the most lucrative years of its life with heavy, repeated shop visits before the engines are retired.

At the same time, their new engine, the LEAP, is replacing the CFM56 at a 1 for 1 rate on aircraft around the world. The LEAP installed base is set to roughly triple between now and the end of the decade.

That’s the beauty of it, they are the market leaders, and they are not losing market share as the hand-off from the previous engines to the new are happening. Spare parts and service revenue on LEAP engines only began flowing to the income statement in 2025.

Safran generated operating income of above €5 billion in 2025, and has guided towards above €7 billion by 2028. The civil aftermarket, the spare-parts business that is the heart of the thesis, grew at double-digit rates last year.

The shares currently trade at €343 which translates to enterprise value of around €142 billion, which is not cheap, giving you a free cashflow yield over 3%.

The market is clearly counting on the current narrative to play out, for the CFM56 base entering its peak overhaul years, the LEAP base tripling by 2030, LEAP spare-parts revenue that only began flowing to profit in 2025, and management’s own guidance lifting recurring operating income.

This is not priced with a steep margin of safety, but nor does it need to be. Your margin of safety comes from the near certainty in the contracted cashflows and its expanding pace.

The stock has been in a structural bull channel since July of 2023, and as it seeks to once again make new all time highs, the path is clear for it to reach €400-500 by next year.

The setup on the chart would be invalidated if it fell to the bottom of this channel, which gives us a stop loss around the €280 level which would need to be raised as months go by to follow up the channel.

Then there is the fact that the USD’s bull market versus the Euro which started in 2009 ended in 2022, and broke below the support last year in 2025.

The chart suggests that the strength in the US dollar is short lived unless it can reclaim 0.9 Euros in short order. If the dollar stays / turns weaker, then a position in Safran benefits naturally from this.

Trade 2: Own the small parts

TransDigm ($TDG) is breaking out of its chart pattern as I type this, so lets lead with the chart. You’re looking at a stock which 8xed since the pandemic, topped at $1600 last year in 2025 and has been consolidating ever since in this falling wedge, of which it is attempting to break out of. A close above $1330 on Friday would validate the breakout, and a break above $1400 would get me more excited.

Here the failure mode is if the stock breaks below $1100, which is a similar 18% decline to the failure mode in Safran.

So that’s the chart, now, what does TransDigm do?

This is a company which has built itself around a ruthless insight: there are many small flight critical parts on an airplane which are proprietary and which have only one legal sources. 90% of Transdigm’s sales come from proprietary products which it designed and owns, and 75% are parts where it is the sole source, with no other company certified to issue them.

As is common in MRO, more than half of its revenue comes from the aftermarket, where it has a high margin replacement business.

The consequence of this is that it has margins which do not resemble that of a manufacturing company whatsoever.

TransDigm earns operating margins before depreciation above fifty percent, a figure that has marched steadily higher for fifteen years as the company has acquired more sole-source positions and pushed pricing through them.

And think about that because it’s the living proof of the benefits of having a hold on a scarce good for which there is strong demand: you can take price and it doesn’t change volumes whatsoever.

Think about what sole-source means for its customers, the airlines. A grounded aircraft might be waiting on a single small valve that only one company is certified to make.

The part might cost a few thousand dollars. The aircraft costs tens of thousands of dollars a day sitting on the ground. The airline will pay almost any price for that part, almost immediately, every time.

Multiply that across tens of thousands of proprietary parts on an aging fleet that is flying more than ever, and the pricing power becomes self-evident.

TDG raised full-year guidance again in May, to earnings before depreciation of about $5.4 billion, up 14% year over year, with organic sales up 9%. This is not the cheapest entry. At 20x earnings and 5x levered, you are paying a full price for a wonderful machine.

You’re paying for quality, and I think that’s fair.

Trade 3: own the contract

The third pick is one of the strangest, but might also be one of the most interesting for the patient investor, as the market seems to have not been able to value it properly whatsoever.

Melrose is a British company, listed in London, which owns the UK’s aerospace champion, which was formerly known as GKN Aerospace.

Buried inside this company is a pile of contracts which very few people seem to understand, called risk-and-revenue-sharing partnerships, or RRSPs.

Decades ago, when engine makers wanted to develop new engines, they needed partners to share the large upfront costs. GKN would put up capital and engineering resources to help the engine maker, and in exchange it would receive contractual rights to a share of the programme’s revenue for the entire life of the programme.

These programmes run around 50 years, the cash invested in the partnerships was spent decades ago. Melrose holds 19 of these partnerships, which touch roughly 70% of the world’s flight hours in one way or another. Of these, 17 are already generating cash, with the last 2 turning cash flow positive by 2028.

The company expects to earn around £22 billion of cash from these contracts over their remaining lives between now and 2050.

This is a pure annuity. They don’t need to build engines, to staff a repair shop, or to manufacture any parts. They hold the rights, and the collect their slice of other programme’s for the next 25 years.

But the market isn’t properly valuing this, as these engine RRSPs are hidden inside a larger duller airframe components business which is more cyclical. The Engines business is a little under half of revenue and about three-quarters of the divisional operating profit.

At about 475 pence per share, Melrose is worth about £5.9 billion in equity, or around £7.3 billion once its £1.4 billion of net debt is added. This is about 11x operating profits, which is well below the two other picks.

Melrose’s whole enterprise value at 475p is about £7.3bn, and that buys three things: the RRSP annuity, the non-RRSP Engines business (OE, repair, government work), and Airframes. Let’s run some back of the napkin numbers.

If we put fair industrial values on the two non-annuity pieces: Airframes earns £156m of operating profit but is cyclical, so let’s call it 9x earnings , or about £1.4bn. Non-RRSP Engines is part of the £520m Engines profit, say £150-200m of it on a better multiple, about £2bn. That is roughly £3.4bn for the rest of the company, which leaves about £3.9bn of the £7.3bn enterprise value sitting on the annuity.

To value that £22bn stream at only £3.9bn, you have to discount it at about 18%. On a sensible breakdown, the market is applying roughly a 15-18% discount rate to a contractual, diversified, revenue-share stream on engines that fly 70% of the world’s hours.

Is that fair? Not, in my view. A fair rate for these cash flows is somewhere around 9-11%. At a 10% discount rate, the annuity alone is worth about £7bn, which is almost the entire current enterprise value by itself, meaning the market is handing you the rest of Melrose close to free.

The stock has been range bound between 400 and 680 pence in a classic rectangle, and is now completing a downward wedge, with an attempt to break to the upside.

If it can close this week above 475, it is worth a starter position.

The failure mode here is at the bottom of the rectangle around 400, which is about 16% below the current price. Similarly the top of the range at 680 must be respected as a target, and also would correspond to the value should the annuity be valued at a 10% discount rate.

The USD has been range bound between 0.7 and 0.84 since 2017, but in April 2025 it broke below the support of the dollars 2009 to 2022’s bull market.

In my view this likely constrains the dollar into the lower half of the range, between 0.7 and 0.77 GBP in the interim. Here again, I am not afraid of owning a foreign stock (funny to call British foreign when I hold the passport) because the setup favours a weaker dollar, unless the dollar can convincingly reclaim 0.8, which I don’t think it will.

Conclusion and notes on position sizing and the use of TA.

I could have written an article on Micron (and I might). I chose for a second article to write instead on a theme which is less discussed, but nonetheless playing out as we speak.

I believe investors should deploy capital where there is the best risk-return, not where there is the most attention.

The setup in the MRO space looks interesting for these names I highlighted today.

But here’s the thing about investing, it’s not as much about the ideas as it is about the discipline.

In this industry, you’re great if you’re right 6 out of 10 times. You can be right 4 out of 10 times and still be great if you have the right discipline.

It’s easy to get married to a narrative, and narratives are dangerous, because they lure you in. Sometimes there is a 1,000 pound Gorilla ready to pounce on your thesis, and you just don’t see him.

Knowing that there are unknown unknowns which could screw us, we should turn to price as a source of truth, and as our ultimate keeper. Price action tells us the behaviour of all buyers and sellers of a security. If the chart develops into a bullish pattern which is supportive of our thesis, then this gives us a reason to enter the position.

If the chart goes against us and forms a bearish pattern, then that is when and where we exit the trade. This doesn’t mean the narrative or the thesis is wrong, but it might. And you’re much better off quitting at a price which stung you than a price which knocked you out.

If the trade is eventually good, then you can come back later when the chart is more constructive.

This becomes great to help sizing positions of bets across your portfolio. The 3 positions listed here have exit points that are between 16-19% below the current share price. That means that if we invest 10% of our portfolios across these names, then we’d lose 1.6% to 1.9% of our assets if the trade goes against us. That’s on the higher end of what I’d like, but gives us good information as to how a trade could expand if we add to it in the future.

In a portfolio, I’d probably initiate these at 2% each, with the option to add to the positions as the chart validates the breakouts and clears further levels.

What’s next

On 15 August 2026 I’ll launch the paid edition of Babylon Burns.The date is the fifty-fifth anniversary of Nixon closing the gold window in 1971, the act that set this whole chain in motion.

We’re still figuring out the exact pricing, but in the interim have set the pledges at low levels we will honor for anyone who decides to pledge now.

I will continue publishing weekly until then. Expect a mix of macro views and ideas sprinkled across.

Two asks before you go. If you want the next letter when it publishes, subscribe, the free work alone will be worth your inbox.

Subscribe now

And if this one was useful, repost it wherever you found it, or send it to one person who you think will like it.

As Babylon burns, we’ll light the cigars.

Share

Sam Kovacs

Babylon Burns is written for informational purposes and not to be construed as investment advice. I write about what I own and how I think. Your decisions are your own; do your own work before acting on anything here.

But the world on which the 20th century’s financial system was built is fundamentally fractured.

I don’t mean to say this in a “sky is always falling” kind of way. It’s simply that there are certain things, namely monetary stability and geopolitical order, which will never be the same.

During the past 50 years, the cost of the things which are the hallmarks of middle class American life have gotten out of control, while the nice to haves have deflated, keeping the perception of tame inflation.

And what’s more, the US government is actively working against its own currency.

In February 2022, the United States and its allies did something never before tried against a nuclear power: they switched off its money. Hundreds of billions in Russian central bank reserves, dollars and euros, sitting in the system, earned and saved like anyone else’s, frozen by decree.

Every central banker outside the Western alliance watched it happen and drew the same conclusion at the same moment: that could be us. The realization that the world’s reserve currency came with strings attached hit hard.

They’ve been acting on it ever since. Central banks have bought gold at double the historical pace for four years running. Foreign central banks now hold more gold than US Treasuries, for the first time since 1996.

Meanwhile, the issuer of the currency they’re walking away from is running a deficit of nearly 6% of GDP, in peacetime, at full employment, with one tax dollar in five going to interest alone.

The United States now pays more to its creditors than for its military.

As you probably know, US debt held by the public is on track to eclipse the WWII record.

The people who manage the world’s money are backing away from the dollar at the same time that America has no choice but to debase it.

Those two themes should dictate investment mandates for the next twenty years.

I’ll be referring to these two themes as debasement and multipolar fracture: the deliberate erosion of money, and the breaking apart of the world order that money was built on.

Everything I will ever publish here is downstream of those two.

Why I’m starting this Letter

I’ve published equity research publicly for eleven years, run a systematic macro hedge fund, and have advised sovereigns on capital and policy.

On 21 February 2026 I sent the report below to a short list of associates and hedge fund clients, which concluded:

The market is not adequately pricing the risk of US military action against Iran. Oil at $66-72 with two carrier groups converging on the region and a presidential ultimatum ticking down represents a significant underpricing of tail risk. The asymmetry favors long energy / defense / shipping.

I left Dubai for Cape Town that week, and avoided the subsequent strife in the Middle East.

I had been going back and forth for the past few years on the idea of starting a global macro newsletter which would synthesize the insights of my living in a dozen countries, advising global south Sovereigns on policy, and my unrelenting obsession with financial markets.

Up until today, I continually deferred.” I’m busy with the fund” ,” I don’t have the time”, “I’m content with my current businesses”, “I’m travelling too much”, “I’m learning how to surf, leave me alone”.

But finally, my good friend George Noble, a former colleague of Peter Lynch’s at Fidelity, convinced me that this would be both valuable to the investment community and for myself.

So this is the first edition, which will set the tone for what’s yet to come.

My frameworks are a mosaic of all the people I’ve read who shaped my thinking, and I’ll name them as we go. Expect the classic sitting on shoulders of giants BS lip service here and there.

Many of these people, unfortunately, are either academics with frameworks but no actionable insights, or they are married to a single asset class.

What I will offer instead, is the distillation of my macro frameworks into an actual active global portfolio, with trades which are shared, sized and time bound.

And one commitment up front: everything recommended in my letters is mirrored with my own capital. I eat my own cooking.

[ SUBSCRIBE BUTTON ]

Debasement

The world you grew up in, one dominant power (the US), one reserve currency (the US dollar), capital going wherever it earned most (mostly US equity) ran so smoothly that we stopped questioning what kept it ticking along.

What kept it ticking along is growth.

Without growth, the capitalist engine would fall apart. And this beautiful engine of ours, really only has three different sources of fuel:: demographics, productivity, and debt.

This is economics 101, derived from Solow’s growth model, and the framework has been modernized and simplified by Raoul Pal at GMI, to whom I owe a debt of gratitude for the elegance of his model.

The West rode population growth and demographics growth for two centuries, and when they sputtered, it had no choice but to choose debt as the growth source of last resort.

But today, even the debt engine is extended, yields less returns than in the past, and is starting to make the whole system lethargic.

Let me explain briefly why every one of these levers has been seized.

Let’s start with demographics. A growing economy needs a growing workforce, and for most of history that was automatic: the United States went from 2.5 million people at independence to over 280 million by 2000, with the share of adults in work climbing from 50% to 67% over the same period..

Then, both growth and participation changed course. Birth rates across the developed world have sat below replacement since the 1970s, so populations are set to shrink, and participation has fallen back toward 62%. In the US in the 1960s, five workers supported every retiree. Today it’s fewer than three. By 2050 it will be two. Japan’s population peaked in 2008, China’s in 2022. Others are following.

Productivity is the second fuel, and it’s been sputtering for fifty years. Between 1870 and 1970, output per worker grew about 2.8% a year, doubling every twenty-five years: a man’s son produced twice what he did.

Then it slowed, even as the technology got more impressive. We built the computer, the internet, the smartphone, and now AI, yet measured productivity growth fell to around 1.5%. Solow’s old line still holds: “you can see the computer age everywhere except in the productivity statistics”. Maybe AI will finally change this. I hope it does. For now, it’s not there in the data, and some research is suggesting that within organizations AI has failed to increase output. My gut is that this is because they are trying to plug AI into archaic workflows rather than rebuild the workflows natively around AI. We might be witnessing something of a productivity paradox whereby productivity declines or stalls at first as we go through a period of upskilling, retooling, and reorganization. Either way, the verdict is that productivity isn’t and hasn’t been what it was between 1870 and 1970.

So governments, households, and businesses pulled the third lever: debt. Debt is future growth, spent now. When it builds something that earns more than it costs, it works. When it papers over the failure of the other two sources of growth, or is allocated to government expenses with no credible sources of return, then we eventually run into trouble. US federal debt has gone from 37% of the economy in 1970 to over 120% today.

The last federal surplus was in 2001, the year Gladiator won Best Picture. It’s only getting worse, as there are no other sources of growth. When you add household and business debt, it’s looking even worse.

When debt gets this big you eventually run into trouble and have to work your way out.

There are exactly four ways out, at least according to Ray Dalio of Bridgewater, who is the man behind the second framework we need to lean on: austerity, default, debasement, or a mix of all three.

You can try austerity: cut spending and raise taxes until the debt falls. At the scale required, that means cutting pensions, healthcare, and benefits that tens of millions of voters depend on. No government that tries it survives the next election, and every politician knows it. Half the constituency won’t take a tax increase, the other half won’t take a benefit decrease. It’s a political non-starter.

You can default: simply refuse to pay. For the issuer of the world’s reserve currency, that isn’t a viable solution. The Treasury bond is the foundation under global trade, banking, and pensions. Choosing to default would be catapulting the US away from the global reserve currency and seriously undermine its position in the global world order. It isn’t happening because unlike Lehman Brothers, Uncle Sam has an ace up his sleeve: Debasement.

The logic is simple. Debt is $39 trillion, and counting. But if a dollar could be worth less than it is now, then the government can repay its debt in the future with funny money that’s not worth what it was when they lent it.

The numbers on your statement still go up while each dollar buys less, and the pain is spread so thin, over so many years, that no one can name the day they were robbed. That’s why it gets chosen again and again by countries in this situation.

It was chosen after the Great Financial Crisis in 2009. It was chosen again in 2020. The US isn’t the first to try this. Rome, Weimar Germany, the Soviet Union before its collapse.

Or you can attempt what Ray Dalio calls a beautiful deleveraging: a careful cocktail of the other three, calibrated to bring debt down without crashing growth. Every government aims for it. In practice it’s like landing a plane with almost no runway left, and even when it works, most of the lift comes from debasement anyway. It’s what we achieved post 2009, and which allowed us to avoid a 1929 style crisis.

So: austerity is refused, default is unthinkable, and the beautiful version still leans on the printing press. Debasement gets chosen again and again, government after government, in every heavily indebted democracy on earth.

Dollars are constantly losing value, and it goes beyond the level measured by inflation. The most sensible way to measure the debasement rate is the gap between M2 and real GDP growth (the mechanics of which we’ll go into more detail in the future).

The fiscal writing is on the wall in the US, in Japan, in China, and in the EU. Currencies have been and will continue to be debased as we attempt to engineer our way out of a debt problem.

This is the backdrop against which investing must happen, and it explains why since the GFC, investing has gotten weird, with valuations moving away from common yardsticks, and old prediction models no longer working.

Multipolar fracture

Outsourcing has been a beautiful source of growth as our participation rates peaked and dwindled. It allowed us to stretch beyond what demographics would imply, and to substitute efficiency gains for geographic arbitrage.

But that dividend has been digested, and has now reversed. The post 1990 equilibrium which brought us globalization and single markets and safe oceans is crumbling one year at a time, one conflict at a time.

Marko Papic and Peter Zeihan have each mapped this terrain from very different angles, and both run all through my thinking here.

Papic’s biggest insight is that we should ignore narratives, and instead focus on the binding political, economic, and geopolitical constraints to assess outcomes.

Zeihan’s biggest insight is that globalization is a temporary blip made possible by the US navy post WWII, but that as demographic stagnation and collapse (the same reason that is killing growth) accelerates, we will transition to a multipolar world.

I think we’re already half way there.

The clearest evidence is in what governments write about themselves. Read the recent national security and economic strategies of the United States, China, Russia, the EU, India, Japan, Iran, Saudi Arabia, the UAE, and Brazil side by side, and they agree on almost nothing, except the shape of the world they’re preparing for.

Even the United States now states plainly that it does not seek permanent global domination (NSS 2025 and NDS 2026 for those who care to check.)

The future is one in which: everyone needs to rearm, new payment rails and reserves get used, new supply chains are organized, and everyone enters a war on resources to fuel their regional economies.

A fractured world builds two or three of everything it used to build once. Multiple chip supply chains, multiple payment systems, munitions industries, etc. That duplication is staggeringly expensive, and expensive means inflationary when it comes from the top down, as it is suggested.

Debasement and multipolar fracture are the consequences of a demographic stagnation, insufficient productivity make-up, and years of pressing the debt button until it made us numb.

Investing as Babylon burns

Babylon was the original empire of accumulation: wealth, debt, grandeur, then decline.

Nothing is new under the sun, and once again it is clear that Babylon is burning.

Warren Buffett and Peter Lynch had it easy compared to you and me.

That means that investments must do two things: keep working whichever way the world splits and fractures, all while outrunning debasement (otherwise you’re just running on a treadmill, going nowhere).

If the currency loses 8% of its real value in a year and your portfolio gains 8 percent, you didn’t make money. Only the return above the rate of debasement is real.

Money has a half-life, which based on my research (which will be published also not so far in the future), the long term trending half-life is fourteen years.

A working life is forty to fifty years long which is three half-lives. Save a dollar at twenty and by your seventieth birthday it buys an eighth of what it did.

There’s only one way to clear a benchmark like that consistently: own things that grow faster than the money shrinks, or that cannot be compressed by printing money.

Cash loses by definition. Bonds pay you back in the very currency being debased. What outruns the fire is ownership of real, compounding assets, and the surest way to find them is to identify the secular trends, the forces so large and so entrenched that they’ll pull capital, revenue, and pricing power toward them for a decade or more, regardless of who wins the next election. Get the secular trend right and time does most of the work for you.

And here’s what makes this tractable instead of a guessing game: the trends flow straight downstream from everything described above. Governments that must debase create demand for assets they can’t print. A world that’s fracturing must rearm, must duplicate its supply chains, must secure energy and materials it used to buy freely. Societies that are ageing must spend on care, while younger ones build their first middle class. None of these are clever bets on what might happen. They’re consequences of what’s already happening, with budgets, laws, and demographics behind them.

The five themes

Follow that logic downstream and the field of what to own narrows to five themes.

Hard money. When governments inflate debts away, own what they cannot print: gold and the miners, silver, Bitcoin, and the few outlier currencies.

Resource scarcity. The world needs far more energy, metal, and materials for electrification, computing, rearmament, and rebuilt factories, after a decade of underinvesting in producing them. Demand is moving faster than supply can answer. Water is a resource. Oil is a resource. Cobalt is a resource. The grid is a resource. Semiconductors are a resource. This theme is defined quite loosely.

Global rearmament. Every major power is rebuilding its military at once, on a scale not seen in fifty years, and adding cyber defence on top. The spending is contracted years ahead and backed by governments. Some of the most beautiful Peter Lynch style growth stocks lie within this.

The demographic dilemma. The ageing that drives debasement also creates two trades: a rich world spending ever more on healthcare and elder care, and a young world forming its first consuming middle class. Both create opportunities of different scales

Restructured rails. The physical and financial plumbing of the world is being rebuilt for a fractured map: new factories, new payment routes, new exchanges, new places to clear capital. The owners of that plumbing get paid whichever side wins.

This was as compressed as I could do a presentation in this article, but I will obviously expand on all of them a lot in coming weeks.

It is by default a wide mandate, because in financial markets, the stories write themselves, and it would be foolhardy to lock ourselves out of too many assets.

Knowing what to own is half the job

Now of course here’s the thing. The gold bugs have been telling you to buy gold for decades. The Bitcoin maxis to go long bitcoin forever. The oil bulls to buy oil. The tech bros tell you to buy semis. The warmongers want you to buy defense stocks.

They are all right at times and all wrong at others. The greatest fallacy as an investor is to get married to an asset class, or worse an individual equity. Because known themes don’t pay out in straight lines.

A theme can be completely correct and still spend eighteen months going nowhere, or down, because it’s colliding with everything else that moves markets: interest rates, liquidity, the business cycle, sentiment.

Gold can spend years giving back gains after getting ahead of itself. Defense stocks can sell off in a recession no matter how full the order books are. Emerging markets can get crushed by a strong dollar even as their consumers multiply. Silver Tsunami investments can sometimes be just yield curve bets disguised, and not pay-off in the interim. The secular trend tells you what to own. It tells you nothing about whether this quarter is the moment to own it.

That’s why my commitment is to 1. not marry an asset class, 2. make positions time-bound with clear failure modes, 3. be brutally honest along the way.

My work will trickle down from the macro (theme, FX, rates), to the micro (fundamentals and price action).

Technical analysis will play the role where it belongs: at the last mile. At the end of the day you can have the best thesis and still lose money to poor timing. Price action is the synthesis of supply and demand for a given asset, and that carries information. My technical analysis is classical Richard Schabacker style, and it is used as a supporting tool once we’ve past an investment thesis through the waterfall of macro to micro tests.

One more thing, and it matters more than most investors realise: the portfolio is global by construction. Own a great business in a currency that strengthens against yours and you get paid twice. A portfolio confined to one country forfeits that entire dimension.

What’s next

On 15 August 2026 I’ll launch the paid edition of Babylon Burns.The date is the fifty-fifth anniversary of Nixon closing the gold window in 1971, the act that set this whole chain in motion.

Each week,there will be a macro read framed through one of the five themes and a deep dive into a single name investment. Behind it, a live global portfolio, concentrated, conviction-driven, every position disclosed, mirrored with my own money, with trade alerts,

Until then, there will be a steady run of free work. Next Letter, I put everything above on one live position for you to get the taste of the work for free: a single name where the theme, the chart, and the catalysts line up at the same time.

Two asks before you go. If you want the next letter when it publishes, subscribe, the free work alone will be worth your inbox.

And if this one was useful,, repost it wherever you found it, or send it to one person who you think will like it.

As Babylon burns, we’ll light the cigars.

Sam Kovacs

Babylon Burns is written for informational purposes and not to be construed as investment advice. I write about what I own and how I think. Your decisions are your own; do your own work before acting on anything here.