Anthropic's IPO filing – and what it could mean for the AI bubble

AI giant Anthropic released the highly anticipated S-1 filing for its IPO yesterday, putting it on track to go public just before Thanksgiving...

A good place to start would be Page 157, where its quarterly numbers are laid out. Overall, they were very impressive... and reinforce why I've never written a negative word about Anthropic. Plus, I'm a happy daily user of Claude.

(Meanwhile, I've been very bearish on ChatGPT maker OpenAI, calling it the greatest cash-burning furnace of all time – archive here.)

Anthropic's revenue is exploding while expenses are rising more slowly, as the company notes on Page 10 of its S-1:

Our compute efficiency is demonstrated by revenue that grew 1,368% year over year for the three months ended June 30, 2026 compared to compute and infrastructure expenses that grew 300% during the same period.

Take a look at this revenue chart posted on X by market strategist Shay Boloor:

As the chart shows, Anthropic generated an operating profit in the second quarter and has burned remarkably little cash. According to Page 10 of its S-1:

As of June 30, 2026, we had raised an aggregate of $141.1 billion of capital and had cash and cash equivalents of $112.4 billion and short-term investments of $19.3 billion.

That means the company has only burned $9.4 billion to create a business that's currently generating annualized run-rate revenues of $77 billion, according to some estimates.

These revenues are being generated by more customers spending more money with Anthropic. According to Page 134 of its S-1:

As more customers adopt Claude and expand their usage, we have rapidly grown our base of customers that generate meaningful revenue. We had more than 11,900 customers with $100,000 or more of revenue on a TTM [trailing-12-month] basis as of September 30, 2026, up from more than 950 as of September 30, 2025. In addition, as customers further embed Claude into their most critical workflows and growth initiatives, many rapidly expand their deployments into higher revenue tiers. As of September 30, 2026, we had more than 1,900 customers with $1 million or more of revenue on a TTM basis and approximately 250 customers with $10 million or more of revenue on a TTM basis.

On Page 10, it outlines its enterprise strategy:

We believe that enterprise customers achieve a compelling return on investment from Claude, as agentic workflows can drive both top-line growth and efficiency. This enables us to expand quickly with customers, as demonstrated by our Net Revenue Retention of approximately 900% as of September 30, 2026. Diffusion also accelerates as our models become more capable and efficient, allowing us to lower the cost per task for customers. For example, Claude Opus 5.5 reduced customer costs per task by approximately 40% compared to Claude Opus 5.

Revenue growth took off a month ago when Anthropic slashed the price of its tokens nearly in half, as this X post by Atreides Management's Gavin Baker shows:

While Anthropic has "only" burned less than $10 billion so far, this number is going to grow dramatically thanks to the massive compute deals it has signed, as this Reuters chart shows:

The deals total a staggering $657 billion, compared with a mere $5.6 billion spent on compute last quarter.

But as noted in this X post by AI investor Dave Lee, the official amount doesn't include:

MORE GOOGLE TPU CHIPS: est. $125-150 [billion], delivered 2028 to 2031
• Depends on capacity, deployment, pricing and other conditions
• Not counted in the filing's commitments table

These huge commitments are the biggest risk for Anthropic. If demand doesn't materialize as quickly as the company projects, it could be in trouble...

In this X post, Box CEO Aaron Levie notes: "The average Fortune 500 customer of Anthropic spends less than .1% of their expenses on Anthropic."

He argues that this means there's lots of room for growth, which is undoubtedly true – but the questions are: How fast will demand increase, and what will happen to pricing?

I think most users will be happy with an AI model that's somewhat less effective than Anthropic's (which is the best on the market) but costs 90% less – like the open-source models from China and elsewhere.

Then there's the risk that Anthropic's "product might end humanity," as the company acknowledges on Page 46 of its S-1:

Concerns have been raised both within our organization and by other researchers, regulators, policymakers, and industry participants that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponized, or allowed to operate with a high degree of autonomy. We have observed that our models are contributing to the pace of our own research and these risks will increase as AI systems substantially automate AI research and development if, for example, they accelerate AI progress beyond our ability to evaluate and safeguard new models, which could adversely affect our business, financial condition, and results of operations.

As an investor, it's hard to evaluate a tail risk that's so unprecedented...

As for valuation, the real-money bettors on Polymarket think that Anthropic is almost certain (96% likely) to reach a $1.5 trillion valuation at some point before the end of the year, 70% likely to hit $2 trillion, and 31% likely to peak at more than $2.5 trillion:

I don't disagree with those numbers. Anthropic is a very impressive, market-leading company that's growing faster than perhaps any business of its size in history. It deserves a huge valuation, though the right number is impossible to know.

As for my prediction that the AI bubble is in the process of bursting, Anthropic's strong filing increases the chances that I'm wrong – or, more likely, simply early. Time will tell...

When all is said and done, only the strongest AI companies will survive – and thrive. My friend Jeff Brown, founder of Brownstone Research, has identified one of these AI winners...

It's a new company behind what he's calling the AI "super chip," which exceeds Nvidia's (NVDA) in every performance category. And this breakthrough could open the opportunity of the decade for investors who get in early.

I'll be joining Jeff to share all the details during a special presentation tomorrow, October 28 – you won't want to miss it. Click here to reserve your spot.

Best regards,

Whitney

P.S. I welcome your feedback – send me an e-mail by clicking here.

P.P.S. Our oldest (30-year-old) daughter, Alison, got engaged on Saturday to her boyfriend, Forrest, in Woodstock, Vermont, not far from where they met at Dartmouth's Tuck School of Business. Congratulations – we're so happy for you! Here are a few pictures:

On Sunday, our younger daughters, Emily and Katharine, hosted a party for them, which included Forrest's parents and my in-laws. Here are pictures (plus one of us when Alison was an infant):