Anthropic's S-1 Explains Why Wall Street is Selling AI

Anthropic remains set on its $2 trillion valuation for an upcoming IPO.

Good investors will pick a side on whether to justify that price tag or challenge it.

Great investors, however…

Choose to negotiate between both, specifically placing probabilities for each outcome and making educated decisions from there.

I’m not talking about whether you should buy Anthropic once it’s public or not.

I’m talking about knowing how likely this company is to bring the rest of the market (if not the financial system) down with it if we see a high probability of $2 trillion being overextended.

Because everything from:

All depend on whether Anthropic is able to go public at this target, and more importantly, sustain it for the coming quarters.

Now here’s my biggest issue with this $2 trillion number they have been throwing around…

Leading companies are able to command the sky-high valuation multiples that Anthropic is shooting for right now (roughly 55x annualized revenue) as long as:

As you will see today, Anthropic is treading awfully close to failing on all three of these marks.

Meaning,

Its 55x multiple leading to a $2 trillion valuation may be challenged in the coming months.

Take growth for example…

Last quarter, Anthropic posted 144% revenue growth only to see that rate fall to less than half at 45% in the most recent quarter.

That’s strike number one.

We’ll get into the other two strikes in this email,

Which give me plenty of reasons to believe this is why Goldman Sachs’ prime brokerage book data unveiled the most aggressive selling of technology stocks in its recorded history.

Let’s get into it,

Anthropic’s S-1 explains why Wall Street is selling AI.

Like I said, we cannot let our bias get the best of us but rather explore the potential for both the bull and bear cases.

Starting with the good of Anthropic, of which there is plenty:

As token prices come down, now 60% off the highs of the year, the cost to deliver on these tokens (compute mostly) also goes down.

All while Anthropic continues to post double-digit revenue growth.

The end result is a recent 10% adjusted operating margin (excluding stock-based compensation.)

We’ll come back to this later, for now know that the business is starting to prove itself somewhat.

Is going from a negative 13% margin to positive 10% enough to justify a $2 trillion valuation so far?

At ~$260 billion on annualized revenue, a 10% margin would come out to roughly $26 billion.

Meaning,

Anthropic would trade at ~77x EBIT for the estimated 2026 financials.

Still sounds a bit rough to me…

Here’s another positive development for Anthropic, which is tied to revenue quality.

Most of their customers are now spending more than $100,000 per year on services, with the second majority spending more than $1 million.

These are most likely medium-sized businesses and even some large enterprises as well, showing a healthy adoption rate and product-market fit.

Now here is where the second strike comes from:

That’s nearly 90% of every dollar Anthropic brings in.

Let’s cover some of the economic mismatch in these receivables before we continue…

Anthropic has to spend roughly $11 in compute infrastructure for every $1 they will make from these receivables, as long as they are able to deliver on the computing capacity being requested.

With heavy concentration among 5 customers, the product-market fit assumption begins to look not as good as we initially thought.

Why?

Because these top customers are the hyperscalers, with the strongest balance sheets in the world and hundreds of billions of liquidity at their disposal.

They don’t really represent Anthropic’s ideal consumer.

Even worse,

These receivables are not locked in, the top customers can cancel or reduce their spend at will!

That’s a huge risk to consider especially as rival models are looking to deliver a very similar service for a fraction of the cost, creating a customer migration concern.

Click here to see where 55% of U.S. companies are sending their AI spend.

One interesting thing I found in the Anthropic financial statements was this:

Where did the ~$70 billion change come from?

They called it a change in the fair value of their convertible securities.

Let me simplify that for you…

Here’s what Anthropic owes its top customers:

For a total $807 billion in total liabilities not declared on the balance sheet.

Why?

These are off balance sheet agreements, and they are not legally required to be reported under GAAP.

Funnily, this is exactly what Enron did, and what all the internet bubble busts were doing to hide how leveraged they really were.

Anyway, that $807 billion represents 59x the company’s contracted revenues.

Here’s my question now:

If Anthropic reached a 10% EBIT margin, and revenue growth is declining,

Then how exactly are they going to make good on this $807 billion number that just keeps getting bigger?

The company somewhat exhausted the amount of convertible instruments it can issue to the market, especially now that its debt has reached junk status.

So financing the compute commitments through debt is no longer an option.

Then there is the receivables revenue, which cannot be counted since it is 100% dependent on first funding the compute commitments.

That leaves you with:

That $213-$253 billion gap will be up to operating cash flow to be closed.

Which was $3.9 billion on an adjusted basis to make for a 10% margin.

Now take this gap into account, because we’re about to use it to check just what the IPO needs in order to hold.

In order to justify this $2 trillion number, Anthropic must generate an estimated $567 billion in annualized revenue.

That’s 100% above the 2026 estimates, and in a forward-looking market, this has to take place in the next 12 months tops.

A more realistic case, based on today’s slowing revenue growth and impaired quality, would be closer to $231 billion.

In other words,

I am placing a higher probability on a valuation of $816 billion as opposed to $2 trillion, which would translate into trillions worth of losses for everyone whose debt and commitments are reliant on Anthropic.

Do you understand why they are racing to IPO as soon as possible?

Google, Amazon, NVIDIA, and neoclouds are all desperate to have a liquid market filled with retail money chasing after the IPO so they can cash out.

Now,

I firmly believe this is quickly becoming the consensus view around the marketplace, and exactly why several hedge funds have dumped their AI risk at a record pace.

Structurally, pretty much all AI baskets are now in either a correction (10% off the highs) or a bear market (20% off the highs.)

Which is exactly why I am so quick to make the assumption of the big AI dumping activity taking place.

In my weekly plans you can see in real time how factors that are opposite to the AI trade have begun to command higher prices while these baskets continue to make lower lows.

Anthropic’s economics are the reason why.

Click here to see where Wall Street is dumping all this risk into.

Here are some of the most important metrics to track in the AI trade positioning:

The combination of these three makes for a very strong signal:

“Smart money” is still very concentrated and positioned in these AI baskets, but their shorts are now just as big (if not bigger) to completely hedge out the long book while they sell.

That’s classic hedge fund behavior…

These people come in and ask:

“How can I get rid of all this without causing a market-wide panic?”

The answer is to short as much as you can out of the opposing view (like software stocks) and begin a controlled orderly liquidation of your long book.

If conviction was high and well in place, there wouldn’t be this much short positioning to hedge while the dumping is done.

How do I know this is software?

Look at the crowding regressions, where all the neoclouds stand on the bottom right while all software stands on the top left.

When this balance tips over, it will be a tug-of-war, a zero-sum game essentially.

One could say this is one of the greatest opportunities of the year if you can play it right.

That’s what I’m looking to do for Offside Premium members toward the end of the year, especially after the Anthropic IPO.

In a few quarters from now, when the failing economics behind AI are clear to the market…

I’m highly confident that this trade will gain popularity.

But,

By then it will already be too late for those who waited.

I remain focused on the memory complex as one of the best ways to express this idea.

Some companies, like Micron, have seen their forward P/Es come down to multi-year lows.

Why?

Markets understand that this industry overshot its investment intensity, and that will lead to oversupply scenarios if demand falls below projections.

Judging by everything we covered today…

I think that’s a very probable scenario coming up.

Still, the timing is far from obvious (as with all bubbles) and remains one of the biggest risks to the memory trade.

So I am approaching it through a long/short options strategy, which deserves the type of attention, guidance, and hand-holding you won’t get in Free Offside newsletters.

I am already in this Micron trade, in fact I opened it in August.

Even after being wrong for nearly three months on the range, I still managed to remain in profit and will even make a decent amount by November expiry.

When you have high conviction on an idea, but are uncertain of the timing, a different strategy needs to be applied.

It’s exactly what we expose you to inside Offside Premium.

That stock that’s up 10% today?

We are also selling cash secured puts on it every month, amplifying our returns as we continue to accumulate a position.

Rotation markets are hard, and I hope you don’t get caught in the tide that tends to wash out pretend traders.

See you on the other side.

Meanwhile, here is an interview with Mohnish Pabrai to get us back into a long-term perspective.

The next decade will be critical to your investment career, as it can either make it or break it depending on where you choose to invest.

This video will help you grab some golden nuggets of information and mental frameworks around this type of longer-term thinking.

And,

If you would like to see how I’m personally hunting for multi-bagger ideas in this new decade scenario…

Click here to see what I found after studying 50 countries and their stock markets.

Until next time,

OFFSIDE RESEARCH

Against the Tape, Ahead of the Curve.