• Anthropic is touting its ability to use different AI chips in its initial public offering filing.
• Analysts say a multichip strategy could weaken Nvidia's pricing power.
• Switching chips is hard and expensive, putting it out of reach for smaller players.

Anthropic is hedging its bets by using multiple chips to handle its AI work — a strategy that could eventually erode Nvidia's dominance.

The AI lab says it manages compute "as a single, fungible pool from multiple strategic partners" — with software that can direct its work to different chips depending on where they're most productive, according to Anthropic's S-1 filing, which came out Monday as the company prepares to go public.

Anthropic says it uses hardware from four chip families: Amazon Trainium, Google TPUs, Nvidia GPUs, and AMD GPUs.

That flexibility, which Anthropic said "took years and significant investment to build," could make major AI companies less dependent on Nvidia over time and give them more negotiating leverage, analysts said.

"The real exposure could be Nvidia's pricing power," said Patrick Moorhead, founder and CEO of Moor Insights & Strategy, "because a customer that can shift work across four platforms negotiates very differently than one that can't."

Ashish Nadkarni, an analyst at IDC, added that a multichip approach also makes Anthropic less vulnerable to fluctuations in Nvidia's chip supply.

Breaking Nvidia's grip isn't easy

Although Anthropic's chip mix isn't new, the company is now pitching it as a core advantage to investors, Moorhead said.

The threat has its limits. Switching between chips is still difficult — especially when it comes to training AI models as opposed to running them, Moorhead said — and only the biggest AI companies have the resources to do it.

Anthropic also continues to spend billions on Nvidia compute, Moorhead said.

Gil Luria, head of technology research at D.A. Davidson, said that while other major companies like OpenAI likely take a multichip approach, it may not be worth the effort and investment for smaller players.

Brad Gastwirth, global head of research and market intelligence at Circular Technology, said Anthropic's strategy doesn't represent a major near-term threat. Having more chip options doesn't erase Nvidia's advantages in performance, networking, and software, he said.

There's a broader risk for Nvidia if AI compute becomes more of a commodity.

Citadel CEO Ken Griffin has warned that if supply overtakes demand, it could drive down rental prices.

"GPUs, like airlines, have a very low marginal cost to run," he told Bloomberg last week. "The moment you have overcapacity, pricing goes into freefall."

Nvidia did not immediately respond to a request for comment from Business Insider.

Have a tip? Contact this reporter via email at gweiss@businessinsider.com or Signal at @geoffweiss.25. Use a personal email address, a non-work WiFi network, and a non-work device; here's our guide to sharing information securely.

• "The Big Short" investor Michael Burry says Anthropic's IPO filing has fueled his worries about AI.
• Burry said a dense web of AI deals has made the industry "catastrophically vulnerable."
• Writing on Substack, Burry described AI financing arrangements as "garish and audacious."

Michael Burry says Anthropic's IPO filing this week has reinforced his concerns that the entire AI ecosystem is at risk of disaster.

The investor of "The Big Short" fame wrote in a Substack post on Tuesday that the first in-depth look at Claude's parent company, ahead of what could be the biggest IPO in history, threw up a bunch of red flags.

Burry pointed to Anthropic's huge spending commitments, slowing revenue, and tangled ties.

He said that Anthropic has struck deals to buy about $537 billion of compute from the likes of Amazon, Alphabet, and Microsoft, and reached preliminary agreements worth hundreds of billions more with others such as SpaceXAI, putting its total commitments at about 12 times its current annual run-rate revenue.

"This is not stated clearly in the S-1, and in my view it is the most deserved headline from it," Burry wrote in a Tuesday Substack post.

He underscored how hyperscalers like Amazon and Alphabet act as Anthropic's "lenders, vendors, channel providers, and shareholders." He also outlined its deal with Broadcom, which involves Apollo and Blackstone and includes convertible notes and structured credit tranches.

"Just remember, the reason this is so complex is because too much debt is needed, and it must be unloaded from the balance sheets of those involved," Burry wrote.

"While some may see all the circularity and connectedness as security on the way up, and that is not wrong, it does leave the entire organism catastrophically vulnerable either on a pause or on the way down," he added.

He also described financing arrangements in the AI space as "garish and audacious," and said hyperscalers aren't generating enough cash to fund the AI buildout so they're resorting to "Byzantine" deals.

Burry said it was a "very similar" situation with vendor financing during the dot-com bubble, saying "people keep making the same mistakes with every new thing."

Burry's worries

In its filing, Anthropic said its revenue surpassed $16 billion for the first six months of this year, a 14-fold increase from the first half of 2025. Its operating loss narrowed by about two-thirds to below $700 million, and it generated nearly $4 billion in operating cash flow.

Burry initially commented on Anthropic's filing in a Substack chat thread on Monday.

He said that revenue was "slowing," citing Anthropic's claim that third-quarter revenue was 86 times the $197 million it generated in the fourth quarter of 2024, or about $16.9 billion.

That would mean Anthropic's revenue grew by about 46% between the second and third quarters of this year, after soaring 144% between the first and second quarters.

"Overall slowing of growth is not attributed to a decline in tokenmaxxing, but I did not expect that to be mentioned," Burry wrote.

Burry was referring to the trend of maximizing AI token consumption, which has fallen out of favor in recent months given its costs.

In another comment, Burry wrote that large language models (LLMs) like Claude are "improving" and "already being used more efficiently and that trend will continue, to the detriment of their growth rate."

Burry said he's focused on the complexity and fragility of the debt financing many AI deals, hurdles to installing new graphics chips shoring up demand for older chips, a potential crash in memory chip prices once the shortage eases, compression and small language models (SLMs) reducing the need for computing power, and the "fundamentals behind not just companies but the deals they are doing."

Responding to a subscriber who said that Big Tech companies are shifting from Anthropic's AI models to their own, and that tokens are rapidly being commoditized, Burry wrote: "Yes, that is what I hear."

Anthropic did not immediately respond to a request for comment on Burry's posts.

The contrarian investor shot to fame after his massive wager against the mid-2000s housing bubble was chronicled in the book and movie "The Big Short." He pivoted from running a hedge fund to writing about his personal investments on Substack last winter.

A recurring theme of his work since then has been deep skepticism of the AI boom. He's flagged slowing growth, overinvestment in microchips and data centers, aggressive accounting, hidden liabilities, and circular financing deals.

The maker of miracles may no longer be God, but it does rhyme with him. 

Anthropic, creator of Claude, filed its S-1 paperwork on Monday in preparation for a gargantuan initial public offering. The document runs over 300 pages and spans more than 240,000 words, some 50,000 more than the New Testament. It asks potential investors to prioritize mission over money: "Mission" appears 79 times, while "money" — the thing an S-1 is filed to generate, in this case, a bajillion dollars worth — a mere 17.

The company professes it could "bend the arc of geopolitical history toward freedom"; it could grow economies everywhere, including "the poorer parts of the world"; it could cure "diseases that have plagued humanity for millennia." Page 4 of the S-1 is not dissimilar from Jesus in Matthew 11: "The blind receive their sight, and the lame walk, the lepers are cleansed, and the deaf hear, the dead are raised up, and the poor have the gospel preached to them."

Anthropic aims to "bring forth an age of miracles," it writes, and the investors who buy in should be true believers. Because in the wrong hands, as you'll see if you turn your missals to page 46, AI can be built in ways that pose "catastrophic or existential risks to humanity."

While S-1s are definitionally visions of the future and full of hype, most are drab financial documents, eye-popping for their prospective profits more than their prose. But Anthropic didn't just open up its book; it wrote its own Good Book, a Bible-like statement of good and evil that joins a growing canon of the company's sacred texts. If you were to stumble onto any random page of this S-1, you'd be forgiven for thinking you were reading not a perfunctory government filing but the scripture of a doomsday religion.

Anthropic has long dodged allegations of fomenting a new, questionable faith. Headlines and takes on Anthropic's cultishness abound: "The Cult of Anthropic" (The Information), "Anthropic is a Cult and Claude is rapidly becoming its God" (The New York Post), "Has a cult ever started a company as successful as Anthropic?" (Peter "Beware the Antichrist" Thiel). Critics point to the employees' extinction predictions — p(doom) is not too far from p(rapture) — and ties to the effective altruism movement. It doesn't help that Dario Amodei was reportedly a Y2K prepper in high school. Anthropic has welcomed the reputation in other ways, inviting religious leaders to oversee the construction of Claude's rules.

In its S-1, Anthropic asks investors to take an actual leap of faith.

Let's play: Is that an Anthropic core value or one of God's 10 commandments?

Anthropic asks its employees to act for the "global good," be "helpful, honest, and harmless," and "hold light and shade." They shall have no other gods — as in, "put the mission first." And, just as followers must sublimate themselves to praise God, Anthropic employees must be "low-ego." The S-1 says it at least nine times.

To understand Anthropic, you'll need to understand 42 ornate terms. The S-1's glossary is sprawling, including neologisms for virtuous acts ("closing the loop") and evil deeds ("prompt injection"). Then there's "Constitutional AI," a type of large language model that is trained with goodness in mind. The SEC's "Plain English Handbook" implores companies to use "concrete, everyday words" and to avoid "relying on glossaries and defined terms." But can you truly be AGI-pilled if you don't speak of "interleaved thinking," "mechanistic interpretability," or "multi-agent orchestration"?

You use language like this because we need to be — and I say we as in society — true believers that this is going to work.James Bort

It's exceedingly hard to get a job at Anthropic, and not just because of the massive pool of cash-eyed engineers seeking IPO windfalls. The company has an intensive interview process, one it name-checks in the S-1. It looks for honesty, trust, and (you guessed it) a low ego. "We are willing to turn down candidates who don't demonstrate them, even if they are extremely strong technically," the S-1 reads. They must, in other words, be devotees through and through.

Accepted engineers become "Ants." Most tech companies have a name for their employees: Googlers, Metamates, Amazonians, Nvidians, Palantirians. Anthropic's moniker evokes the image of a collective, company as colony. In Proverbs 6:6, ants are considered diligent and prepared. Staffers may yell its teachings at their OpenAI foes: "Go to the ant, you sluggard; consider its ways and be wise!"

CEO Dario Amodei — Amodeity — leads the ants' march, along with his six apostles. These cofounders pull equal cash salaries and are "distinctly equipped to be stewards of our mission," the S-1 says. And, while the company opens itself to a broader group of believers through its forthcoming IPO, the divine cofounders will retain majority decision-making power.

Anthropic has its own form of tithing: The cofounders all made voluntary pledges of 80% of their personal equity to philanthropy. They incentivize their employees to do the same, promising to match donations.

Cult is an overused word, and many parts of Anthropic are far from cultish. Where cults are led by a liturgical leviathan that squashes the speech of their participants, Anthropic touts its culture of open dialogue. "Our teams operate in a truth-seeking environment where the best argument wins, not the loudest voice or the most senior title," the S-1 reads.

Americans' faith in the presidency, Congress, the Supreme Court, the military, the police, higher education, public education, news media, the medical system, and traditional organized religion are each at record- or near-record lows — and just as AI has brought on tremendous economic and social uncertainty. In that cauldron of doubt, tech founders — with their world-historic money, power, and ubiquity, and in their pursuit of it — sense an opportunity, to offer promise and purpose. As Sam Altman once said, "The most successful founders do not set out to create companies. They are on a mission to create something closer to a religion." They may also be trying to build devotion out of necessity, as the economy and stock market increasingly relies on AI gains.

"If these companies fail, what happens?" asked James Bort, an assistant professor of entrepreneurship at DePaul University, and coauthor of a recent research paper on "cultic startups." "You use language like this because we need to be — and I say we as in society — true believers that this is going to work."

Anthropic didn't respond to my request to know more about how it wrote the S-1. So, we asked Claude: If the S-1 is a confession of faith, what are you? I am its product, Claude said. "The word made flesh, or at least made weights."

• Anthropic shows strong profit potential, with revenue growing way faster than expenses.
• Anthropic's revenue per employee exceeds $6.6 million, outpacing estimates for Meta and Google.
• Anthropic has a surprisingly diverse customer base and huge cash reserves.

For years, one of the biggest unanswered questions about the artificial intelligence boom has been surprisingly simple: Can the companies building the world's most powerful AI models actually make money?

After spending hours poring over Anthropic's IPO filing on Monday, we finally have an answer: Yes. And potentially lots of it.

A couple of Chinese AI labs went public earlier, revealing appalling financial results that sowed doubt about the industry's economics. Now Anthropic has provided a much more compelling demonstration of how this business can work at enormous scale.

The numbers are remarkable. Anthropic generated $11.6 billion in revenue during the second quarter of 2026, up 1,368% from a year earlier. During that same period, spending on computing and infrastructure increased 300%, while other expenses rose 741%

So, the company is bringing in money much faster than its most important costs are rising.

For every dollar Anthropic spent on computing in the second quarter, it generated $2.08 in revenue, up from just 40 cents in the first quarter of 2025. That's a more than fivefold improvement over five quarters.

The result: Anthropic recorded $974 million in quarterly operating profit, representing an 8% margin. Its adjusted operating profit margin reached 19%, compared with -129% a year earlier.

"Frontier-lab unit economics can turn positive at scale," Mark Mahaney, a top tech analyst at Evercore ISI, wrote in a research note on Monday.

Better than Meta and Google

There's more evidence that Anthropic has built an unusually efficient business.

Over the year ending September 30, revenue increased roughly 1,200%, while its employee count grew about 170%. Revenue per employee jumped from $1.4 million to more than $6.6 million.

For comparison, Mahaney estimated that Meta generates about $3.2 million in annual revenue per employee, while Google generates about $2.5 million.

That comparison isn't perfect. Anthropic relies heavily on cloud providers, whose employees handle much of the computing infrastructure and distribution work. The improvement is still striking.

Customer surprise

Another surprise: Anthropic doesn't depend as much on a handful of enormous customers, as some investors had feared.

The company had more than 11,900 customers generating at least $100,000 in annual revenue as of September, compared with over 950 a year earlier. More than 1,900 customers spend at least $1 million annually.

No single customer accounted for more than 10% of revenue during the first nine months of 2026. Its five largest customers represented about 21% of third-quarter revenue, down from roughly 36% in the first quarter.

"Concentration may be better than feared," Mahaney wrote.

Cash cushion

Anthropic has a formidable financial cushion. Its June 30 balance sheet showed more than $130 billion in cash and short-term investments.

And Mahaney estimated that the company has burned through just $9 billion since it started more than five years ago. That's a long runway.

Now, don't ask me whether all this makes Anthropic worth $2 trillion, $1.5 trillion, $1 trillion, or considerably less. If I knew that, I'd be making a lot more money as an investor. And even professional investors don't know.

An $800 billion caveat

Some enormous caveats: Much of Anthropic's revenue comes from customers paying for what they use, rather than signing long-term contracts. Meanwhile, some of its computing agreements require minimum payments whether or not it needs all the capacity.

Mahaney estimates that Anthropic has entered into nearly $800 billion in computing arrangements, although only about $266 billion was fixed or guaranteed as of June.

That's still a staggering amount.

"The commitments assume revenue compounding that is not contracted," Mahaney wrote.

Translation: While the company has committed to spending hundreds of billions of dollars on AI computing infrastructure, it hasn't secured the future revenue needed to pay for all of it.

Importantly, most of the potential $800 billion exposure isn't fixed. Many arrangements depend on capacity being delivered, can be adjusted, or include cancellation provisions.

In-house compute and capex

Another wrinkle: Anthropic warned Monday that its spending will increase as it shifts toward leasing AI chips and operating more of its own computing infrastructure, rather than relying entirely on Amazon, Google, and other cloud providers.

That could make future cash flow considerably less attractive, even if operating profits continue to improve.

Still, the central takeaway from Anthropic's IPO filing is difficult to ignore.

For years, investors have wondered whether selling AI could become a genuinely profitable business, rather than an extraordinarily expensive scientific experiment.

Anthropic has now shown that it can. The next question is whether it can keep doing so as the bills get much, much bigger.

Sign up for the AI Insider newsletter here. Reach out to me via email at abarr@businessinsider.com