Anthropic’s IPO Filing Reveals AI’s Biggest Financial High-Wire Act

Anthropic has put a giant number on the AI boom: $974 million in operating profit for the second quarter of 2026. The Claude developer’s IPO filing also shows explosive sales, enormous cash reserves, and infrastructure commitments that could reshape the company’s future.

Filed with the US Securities and Exchange Commission on October 26, the registration statement opens the door to a public listing on Nasdaq under the ticker ANT. The document arrived around 9 AM on October 26, and Anthropic can begin its roadshow in 15 days.

A Financial Breakthrough Before the Market Debut

Anthropic reported $11.6 billion in revenue for Q2 2026, alongside its $974 million operating profit. Revenue rose 1,368% year-on-year, giving the company a striking financial profile as it prepares to enter public markets.

For the first half of 2026, the company generated $16.3 billion in revenue. The filing materials also describe revenue rising 13 times, from $1.3 billion to $17 billion over the same period last year. Together, those figures show the speed of Anthropic’s expansion, even as the reported totals reflect different figures presented in the IPO materials.

The turnaround looks just as dramatic in operating results. Anthropic recorded a $693 million operating loss for the first half of 2025, but adjusted operating income reached $1.62 billion during the first six months of 2026, compared with a $1.85 billion loss in the year-earlier period.

Another set of figures puts adjusted operating income at $1.6 billion, producing a margin above 9%. Anthropic also collected $3.93 billion in cash from operations, spent less than $500 million on capital expenditures, and registered around $3.5 billion in positive free cash flow.

The enterprise engine is driving that surge. Claude Code has helped expand adoption, and more than half of Fortune 500 companies were customers that generated at least $100,000 in revenue over the 12 months ended September 30. The S-1 also describes a base of 11,000 enterprises producing more than $100,000 a year, up from 1,500 last year.

Cash Rich, Compute Hungry

Anthropic held $131.65 billion in cash, cash equivalents, and short-term investments as of June 30. Another figure in the filing materials places cash and short-term investments at $122 billion after a $65 billion funding round in late May.

The company’s balance sheet matters because its business consumes extraordinary amounts of computing power. Compute and infrastructure expenses reached $8.95 billion during the first half, making them Anthropic’s largest operating expense.

In Q1, sales exceeded AI infrastructure costs by 40%. In Q2, the gap widened: Anthropic reported $11.6 billion in revenue against $5.6 billion for compute. That spread gives the company room to grow, but the filing makes clear that future capacity will demand far more money.

Anthropic has committed to more than $600 billion in computing and infrastructure spending across approximately eight partners over the coming decade. The commitments include at least $110 billion with Amazon, $114.5 billion with Google, and $44.9 billion with Microsoft.

The S-1 details seven contracts under infrastructure service agreements and infrastructure equipment leases, with a combined value of $630 billion. Seventy percent of those arrangements, measured by dollar value, are non-cancellable, and approximately $441 billion represents guaranteed commitments.

Some agreements require payments even when Anthropic does not use the reserved compute. Most contracts appear to run for seven years or less, creating at least $60 billion in annual obligations that sales must cover. Anthropic also plans to buy chips in some cases and assume the cost of equipping dedicated centers operated by others.

The company stated its intention “to secure further compute beyond our existing commitments, [and] expand our own data center footprint to secure additional long-term compute.” That plan points toward a larger infrastructure budget ahead, not a smaller one.

The IPO’s Huge Promise and Real Risks

Anthropic’s workforce stood at approximately 5,300 employees as of September 30. Based on that headcount, first-half revenue per employee was about $3.1 million, or $6.1 million annualised, using the reported workforce figure rather than the period’s average headcount.

Under GAAP, the company reported an $80.7 billion net loss for the first half of 2026, also described as an $80 billion loss. The result was driven largely by a $72.4 billion accounting charge tied to changes in the fair value of financing instruments. Adjusted results and cash flow tell a different story, but public investors will see both pictures.

Investors predict a valuation between nearly $1.5 trillion and $2 trillion, with one anticipated range set at $1.8 trillion to $2.0 trillion. At that level, Anthropic would rank second only to SpaceX as the largest market debut of all time; SpaceX raised about $86 billion and reached a valuation of about $2 trillion in June.

The IPO could amass another $100 billion, lifting Anthropic’s reserves to well over $200 billion by late this year. The filing does not reveal the expected starting stock value, so the final valuation remains open.

Anthropic warns that revenue growth may slow, infrastructure spending may rise, and dependence on a limited number of cloud providers creates operational risks. The prospectus also lists model misuse, security vulnerabilities, and the possibility that AI capabilities could advance faster than the company’s safety measures.

One unusual provision could affect future shareholders: the board can issue an additional 2% of outstanding shares to the Anthropic Public Benefit Fund each time the stock rises 50% over a benchmark. If the stock climbs, each shareholder owns 2% less of the business; if the price later returns to its starting point, shareholders are worse off.

Now Anthropic is moving from private AI powerhouse to public-market test case. Its numbers show an enterprise business with extraordinary momentum, but its commitments show the price of staying at the frontier: every leap in Claude’s reach demands another leap in compute, capital, and accountability.

Stay connected via Google News

Follow us for the latest AI News