Anthropic Is Spending Nearly $900 Billion on AI. Eventually, Wall Street Will Demand Profits.

Artificial intelligence companies have spent the past several years convincing investors that AI will change the world. Now they're getting a chance to prove they can make money doing it.

Anthropic (ANT), the company behind the Claude AI models, disclosed in a regulatory filing Monday that it has committed to spending approximately $895.4 billion on computing and infrastructure. That's an extraordinary amount of money for a company preparing to go public, even one experiencing explosive revenue growth.

And while Anthropic has some impressive financial results to show for its efforts, that enormous spending commitment raises a question that investors shouldn't ignore.

Can the company generate enough cash to pay for its AI ambitions?

Anthropic’s Massive Computing Bill

According to Anthropic's amended S-1 filing with the Securities and Exchange Commission, the company's obligations include hardware leases, cloud computing agreements, and other infrastructure commitments stretching as far as 2045.

Broadcom (AVGO) accounts for $161.2 billion in equipment lease payments, excluding interest. Alphabet's (GOOG) (GOOGL) Google has another $121.3 billion in minimum cloud commitments through 2033.

But there's more. Anthropic estimates it could spend an additional $125 billion to $150 billion on Google tensor processing unit (TPU) computing capacity, an amount that isn't included in the contractual obligations table.

The list of other technology partners reads like a who's who of the AI industry.

Anthropic has committed $110 billion to Amazon (AMZN) and at least $44.9 billion to Microsoft (MSFT). It also has agreements worth up to $120.5 billion with SpaceX (SPCX), up to $44.6 billion with Nscale Power (SMR), and approximately $41.3 billion with privately held Lambda. Anthropic says about $213.8 billion of its commitments include termination rights or conditions that must be satisfied before arrangements commence.

Still, the payment schedule is daunting. It’s expected to incur $447.5 billion by the end of 2030, including $69.8 billion by Sept. 30, 2027.

Revenue Growth Is Extraordinary

To be fair, Anthropic isn't spending this money without generating substantial business.

Revenue in the third quarter was $17 billion, more than 12 times its revenue from the same period last year. The company earned $33.2 billion in the first nine months of the year versus $2.5 billion in the first three quarters of 2025.

Enterprise adoption is growing rapidly, too. Anthropic said it got more than $100,000 in revenue from more than 11,900 customers in the past 12 months, including approximately 250 customers who spent at least $10 million. That's compelling evidence that businesses are finding real value in Claude.

And Anthropic's operating performance has improved considerably. The company posted an operating loss of $1.3 billion in the first nine months of 2026, an improvement from a year ago when it lost $5.2 billion in the first nine months of the year. It generated $6.6 billion in operating cash flow during that period, reversing a $2 billion outflow in 2025.

The third quarter produced a $567 million operating loss after a profitable second quarter, partly because Anthropic relied on expensive short-term computing capacity.

The cost of providing AI services remains a serious consideration.

Yet Wall Street Will Want Proof

I've previously written about the “AI Red Queen Effect”—the pressure on technology companies to keep spending enormous sums simply to remain competitive.

Anthropic is a particularly interesting example. It needs computing capacity to develop better models and serve more customers. But those investments create financial obligations years before the company knows how much revenue the equipment will generate.

The company isn't short of money today. Anthropic reported $146.6 billion in liquidity at the end of September, including $95.6 billion in cash and cash equivalents.

But nearly $900 billion in commitments dwarfs that financial cushion. My concern is what happens when Anthropic becomes a publicly traded company. Investors may initially celebrate its extraordinary revenue growth, but enthusiasm alone won't support its valuation indefinitely.

Eventually, shareholders will want evidence that Anthropic can consistently generate profits and enough cash to fund its expansion without relying on repeated capital raises. And if it can’t provide that proof, the market will punish Anthropic's stock.

Anthropic has already demonstrated that businesses will pay billions for AI. Now it needs to demonstrate that providing those services can become a sustainably profitable business. Until then, I expect the company's enormous spending commitments to put pressure on its shares, regardless of how exciting its technology becomes.

On the date of publication, Patrick Sanders did not directly have a position in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.