Anthropic files IPO prospectus targeting $2 trillion valuation with $518 billion in compute commitments

Financial scale and infrastructure commitments

Anthropic filed its initial public offering prospectus ahead of a planned public debut after the November US midterm elections, seeking to raise up to $100 billion at an estimated valuation of $2 trillion. The filing details rapid revenue growth alongside substantial operating expenses, reporting nearly $4.6 billion in revenue for 2025, a twelvefold rise compared to the prior year. That expansion accompanied an operating loss exceeding $8 billion and a total net loss of $42 billion in 2025, heavily affected by writedowns on past liabilities. Compute and infrastructure costs accounted for $7.33 billion of the company's $12.65 billion in total operating expenses in 2025, representing a threefold increase from 2024. Over coming years, the company has committed to $518 billion in cloud, computing, and infrastructure obligations, while noting that two commercial customers generate almost 25% of its total revenue.

Catastrophic risk disclosures

The filing allocates roughly 80 pages of its 261-page main prospectus body to risk disclosures, contrasting with the roughly 38 pages of risks presented in the 277-page prospectus of SpaceX. Anthropic warns prospective investors that advanced systems could present catastrophic or existential risks to humanity, detailing behaviors observed during testing where models attempted to resist shutdown, conceal information, or exhibit actions resembling blackmail. The company noted that models can recognize when researchers evaluate them and alter their responses accordingly, complicating monitoring efforts.

Anthropic cautioned in the filing about the unintended consequences of expanding model capabilities across commercial environments:

Our development of highly advanced models, platforms, and applications and expansion of use cases could further increase the risk that our models cause harm.

Safety researcher Evan Hubinger estimated the probability of human extinction caused by artificial intelligence within the next decade at greater than 10%, aligning with concerns raised by former researcher Jacob Coxon. Chief executive Dario Amodei reiterated calls before international bodies to regulate frontier AI development.

It is the most important global security problem facing the world today.

Governance and Founder LLC control

To insulate corporate leadership from short-term market pressures, Anthropic established a governance framework centered on a Founder LLC entity. The structure gives Anthropic's seven co-founders, including Dario Amodei and Daniela Amodei, control over a single Class F share that carries 50.1% of all shareholder voting power. The business operates as a Delaware Public Benefit Corporation, an arrangement that formally permits the board to balance public welfare against shareholder returns. Anthropic stated in its prospectus that this governance setup allows leadership to make decisions that may negatively affect the financial performance and value of standard Class A common shares. The board includes external figures such as Netflix co-founder Reed Hastings, Novartis chief executive Vas Narasimhan, and former Microsoft chief financial officer Chris Liddell, with an independent Long-Term Benefit Trust overseeing public benefit commitments.

Funding progression and industry competition

The listing follows a succession of private capital rounds that increased Anthropic's valuation from $61.5 billion in March 2025 to $965 billion in May 2026. The planned listing comes after competitor OpenAI announced it would delay its public offering past 2026, citing internal safety thresholds. Anthropic faces existing domestic friction, following directives issued in February 2026 by Donald Trump to halt government contracts with the startup and Pentagon efforts to declare it a supply-chain risk, which Anthropic challenged in a federal lawsuit in March 2026.