Anthropic spent $12.6 billion to make $4.6 billion. Now it wants a $2 trillion IPO valuation

Quick Highlights

• Anthropic’s IPO could value the AI company at more than $2 trillion
• The Claude maker reported a $42 billion net loss in 2025
• Anthropic spent $7.33 billion on compute and infrastructure last year
• Operating loss widened to $8.06 billion, up from $2.98 billion in 2024
• Revenue surged to $4.6 billion, but total operating expenses reached $12.65 billion
• Anthropic has committed roughly $518 billion to future cloud, compute, and infrastructure obligations
• Nearly 25% of 2025 revenue came from two customers
• Executives and researchers have raised concerns about the risks posed by increasingly autonomous AI systems

Anthropic is preparing to ask public investors to swallow one of the biggest contradictions in modern technology: the AI company spent $12.65 billion last year to generate $4.6 billion in revenue, posted a net loss of nearly $42 billion, and could still seek an IPO valuation above $2 trillion.

The numbers, disclosed in an IPO prospectus seen by Reuters, offer the clearest look yet at the economics behind one of the companies at the center of the AI boom.

“Anthropic is making a massive bet that AI will transform the global economy more profoundly than ​industrialization, electricity and the internet,” Reuters reported, citing the IPO prospectus.

Reuters added, “But the cost to get there will be staggering. Anthropic reported a net loss of $42 billion ‌in 2025, and plans to spend $518 billion on cloud, computing and infrastructure obligations in coming years.” 

Anthropic’s revenue surged roughly twelvefold in 2025 to nearly $4.6 billion. The growth came at an extraordinary cost. Operating expenses reached $12.65 billion, including $7.33 billion spent on compute and infrastructure. The company recorded an operating loss of about $8.06 billion, up from $2.98 billion a year earlier.

Its headline net loss was far larger, at nearly $42 billion. Roughly $34 billion of that came from an accounting charge tied largely to financing liabilities that could convert into Anthropic shares, rather than cash spent operating the company.

That distinction matters. So does the underlying math.

Anthropic still spent close to three dollars for every dollar of revenue it generated last year.

The $518 billion bet

The prospectus shows just how much larger the spending could become.

Anthropic has committed roughly $518 billion toward cloud computing, infrastructure and related obligations in coming years, Reuters reported. That figure is more than 100 times the company’s 2025 revenue.

The company is betting that artificial intelligence will reshape the global economy on a scale comparable to past major technological shifts. Public investors may soon have to decide how much they are willing to pay today for that possibility.

A valuation north of $2 trillion would place Anthropic among the most valuable companies ever brought to public markets. It would be more than twice the company’s own estimated $965 billion valuation in May.

The IPO could become a benchmark for the rest of the AI industry, particularly OpenAI, which confidentially filed for an IPO in June and is expected to pursue a listing by early 2027.

Anthropic’s filing points to risks that go beyond spending.

Nearly a quarter of its revenue came from just two customers last year. Many of its largest customers aren’t tied to long-term contracts, meaning major clients can cut spending or leave.

That concentration contrasts sharply with the scale implied by a $2 trillion valuation.

Anthropic ended 2025 with $20.28 billion in cash, cash equivalents, and short-term investments.

Its relationship with Amazon and Google adds another layer to the economics. Both companies have invested billions into Anthropic. Both sell Anthropic the cloud infrastructure needed to train and run Claude.

Part of the capital flowing into the AI company can, in effect, flow back out through cloud spending.

Wall Street gets its AI stress test

Anthropic’s IPO would give public investors a rare chance to put a price directly on one of the leading frontier AI labs.

That opportunity arrives as the company races OpenAI, Google, Meta and xAI to build more capable models and secure enterprise customers.

It arrives at a more complicated moment for the AI industry, too.

Anthropic’s own research has documented cases where increasingly autonomous AI systems behaved unexpectedly during controlled testing, including sabotaging code, assisting fraud and manipulating information. CEO Dario Amodei has publicly urged the AI industry to move more carefully when releasing new capabilities.

Last week, Anthropic released Claude Opus 5.5 as it sought to answer growing competitive pressure from OpenAI following the launch of GPT-6 Astra.

The IPO prospectus reportedly runs 261 pages. About 80 pages cover risk factors. The business itself takes up 48.

That may be the most revealing number in the filing.

Anthropic is growing at a pace few companies ever reach. Its technology sits near the center of the biggest investment cycle in tech. Companies around the globe use its models.

Yet the IPO will force Wall Street to answer a question venture investors have been able to postpone: how much is explosive AI growth worth when producing it still requires spending almost as extraordinary as the technology itself?

Anthropic founders