Anthropic Files for IPO at up to $2 Trillion Valuation Despite $80.7 Billion Loss and Exploding AI Costs
The AI company reported surging enterprise revenue but disclosed infrastructure commitments, operating losses and founder voting control

Anthropic has filed a 335-page Form S-1 prospectus for a Nasdaq listing under the ticker ANT, with prospective investors reportedly discussing a valuation of $1.8 trillion (£1.35 trillion) to $2 trillion (£1.5 trillion). Private investor estimates have placed its potential value between $1.5 trillion (£1.13 trillion) and $2 trillion (£1.5 trillion).
The company has not disclosed the proposed IPO's size or price range in its S-1 filing. However, various reports claim that Anthropic aims to raise up to $100 billion (£75 billion) to cover cash shortfalls and fund long-term infrastructure obligations. The listing could eclipse SpaceX's previous record IPO.
Revenue Jumps 14-Fold as Enterprise Demand Grows
Anthropic reported revenue of $16.3 billion (£12.2 billion) in the first half of 2026, a 14-fold increase from the same period last year. Second-quarter revenue reached $11.6 billion (£8.7 billion), driven largely by developer and enterprise demand for Claude Code and business deployments of its technology.
Despite that growth, the filing shows an operating loss of $693 million (£520 million) for the first half, down from $2.1 billion (£1.6 billion) a year earlier.
Anthropic reported a net loss of $80.7 billion (£60.5 billion), including $72.4 billion (£54.3 billion) in non-cash fair-value accounting adjustments. The charges reflect changes in the fair value of convertible notes, derivative liabilities and Simple Agreements for Future Equity (SAFEs).
As Anthropic's valuation increased, the value of those financial instruments rose. Accounting rules require the increase to be recorded as an expense, despite no cash leaving the company.
AI Infrastructure Commitments Reach Hundreds of Billions
Anthropic's disclosed computing commitments include approximately $114.5 billion (£85.9 billion) with Google Cloud through July 2033, $110 billion (£82.5 billion) with Amazon Web Services through April 2036, $44.9 billion (£33.7 billion) with Microsoft Azure through May 2033 and up to $120.5 billion (£90.4 billion) with SpaceX.
These agreements have different contractual periods and terms.
Anthropic's annualised spending commitments across hardware and cloud partners are reported to range from $45 billion (£33.75 billion) to $66 billion (£49.5 billion). Multiplying that range by ten would produce an illustrative estimate of $450 billion (£337.5 billion) to $660 billion (£495 billion), rather than a disclosed ten-year contractual total.
The prospectus identifies reliance on a small group of cloud and chip suppliers as a business vulnerability.
Founder structure limits shareholder influence
Anthropic is classified as a controlled company under Nasdaq rules. A single Class F share held by Founder LLC carries 50.1 per cent of voting power, while the Long-Term Benefit Trust elects a majority of the board.
The filing says public investors cannot bypass management or the trust, even if they collectively favour a takeover. That structure could prevent shareholders from receiving an acquisition premium.
The seven co-founders each hold between 230 million and 246 million Class A shares. The document devotes 93 pages to risk disclosures, including the possibility that future models could exhibit 'self-preserving behaviours', assist in developing dangerous weapons or pose catastrophic risks.
Pentagon Dispute Remains Under Appeal
The filing also addresses Anthropic's dispute with the US government. President Donald Trump and Secretary of War Pete Hegseth designated the company a national security supply chain risk after it refused to allow unrestricted military use of its technology.
Anthropic sued the Pentagon, alleging retaliation. A California federal judge ruled in the company's favour, while a Washington, DC, federal appeals panel sided with the government.
The government appealed the California ruling, and Anthropic sought a rehearing of the Washington decision. Hours after the designation, OpenAI secured its own Pentagon deal.
© Copyright IBTimes 2026. All rights reserved.

