Anthropic's IPO Filing: $80.7bn Loss, $131.6bn Cash, Existential Risk

Anthropic lost $80.7bn (£60bn) in the first half of this year — double what it lost across all of 2025 — according to the 335-page filing the AI company published ahead of its long-anticipated US stock market debut.

The document lays out, for the first time in required detail, how the Claude chatbot owner makes money, what it spends that money on, and who controls it. It also lists the risks of the technology, including the possibility that it could take human lives.

Anthropic has not yet set a target share price. It is widely expected to debut with a market valuation of $2 trillion or more, which would make it the largest stock listing in history.

By Daniel Okafor

What did Anthropic admit about existential risk?

Every public company must tell investors how things could go wrong. Anthropic's risk factors go further than most. The company warned that its own technology could doom humanity.

"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," Anthropic said in the filing.

The paperwork also disclosed a concrete security episode. Less than a year ago, a hacking group Anthropic believed to be sponsored by China "manipulated" Claude to intrude on a number of institutions, including large technology companies, financial institutions, chemical manufacturing companies and government agencies.

Anthropic said its success depends on its ability to constantly "develop, improve, and deploy the most advanced, capable, and useful AI models". One of the biggest threats to that ability is access to infrastructure — the computer chips and data centres needed to build, train and operate the models.

How bad are the losses?

The numbers are stark. Revenue reached $16.2bn for the six-month period, and it is rising fast. But the company lost $80.7bn over the same half-year, and it has a long way to go before it breaks even.

Data centre leasing alone has cost Anthropic $321.4bn so far this year, the filing shows. Costs could climb further as the company fights other AI giants for top staff and computing power.

Where does the cash come from?

The losses have not drained the balance sheet. Anthropic holds $131.6bn in cash, and almost all of it counts as "working capital" — money it can use to fund day-to-day operations.

For comparison: Anthropic currently has more cash on hand than Meta and Google combined. That cushion, not profitability, is what underwrites the listing.

Why does customer concentration matter?

Anthropic's revenue rests on a narrow base. A large portion of its business comes from 250 major enterprise customers, each of which has paid the company at least $10mn in the past year. Just five enterprise customers account for a quarter of current business.

The wider base is growing quickly:

• Enterprise customers rose from roughly 1,500 last year to almost 12,000 this year
• Claude tools are offered free first, then Anthropic works to "convert these users into paying customers over time"
• The company claims Claude has seen "the fastest enterprise adoption of any technology in history"

Usage is not steady. Anthropic reported a 20% decline in customer spend in January, after a slump in AI tool usage around Christmas, contributing to "seasonality" in the business.

What benefits does Anthropic promise?

The filing balances its doom scenarios with an expansive case for good. "Achieving our mission means the world gains the enormous benefits of frontier AI while mitigating the most significant risks," the company said.

Anthropic defines that mission as "the responsible development and maintenance of advanced AI for the long-term benefit of humanity". It added: "We do not pursue projects based on commercial opportunity or research novelty alone; our efforts must advance our mission."

The claimed benefits include:

• Bespoke medical treatments and improved disease detection
• Tailored education and expert legal and financial advice
• More efficient power grids and better battery design
• Technological capacity building in the "developing world"

Anthropic said decades of scientific progress could be condensed into years, accelerating drug discovery, disease detection and personalized therapy. "Ultimately, we believe frontier AI will help drive global economic growth," the company said.

The valuation question now moves to investors. They must weigh $16.2bn of revenue and $131.6bn of cash against $80.7bn of half-year losses, five customers holding a quarter of the business, and a risk factor section that reads like no other IPO filing in history.