Anthropic Files for Nasdaq IPO, Reveals $16.3 Billion Revenue and AI Existential Risks

Anthropic Prepares for Nasdaq Debut

Anthropic, the prominent developer behind the Claude AI model and chatbot, has officially filed its S-1 prospectus with the U.S. Securities and Exchange Commission (SEC) for an Initial Public Offering (IPO). The company plans to list its Class A shares on the Nasdaq Global Select Market under the ticker symbol “ANT.” This move positions Anthropic for one of the technology industry’s potentially largest stock market debuts, with a valuation possibly reaching up to $2 trillion, potentially surpassing SpaceX’s record.

Financials and Massive Compute Commitments

The S-1 filing reveals significant financial details for Anthropic. For the first half of 2026, the company generated $16.3 billion in revenue, a sharp increase from approximately $1.2 billion in the same period last year. However, Anthropic also reported a net loss of $80.7 billion. A substantial portion of this loss, $72.4 billion, is an accounting adjustment related to the company’s soaring valuation, not actual cash expenditure. The operating loss was considerably smaller at $693 million, down from $2.1 billion a year prior. On an adjusted basis, Anthropic posted a $1.6 billion operating profit in the first half of 2026, achieving a margin of around 10%.

A major component of Anthropic’s expenditure is its compute infrastructure. The company has enormous obligations under compute and infrastructure agreements, potentially totaling between $782 billion and $807 billion over the next decade. These commitments include deals with major providers such as Microsoft, Google, Amazon, and Broadcom, many of which are non-cancellable and dictate payment regardless of usage. For instance, an agreement with SpaceXAI for up to $120.5 billion through 2029 includes a 90-day cancellation clause for either party.

AI Risks and Unique Governance Structure

Anthropic’s prospectus dedicates 94 pages to risk factors. The company warns that its highly advanced AI models could pose “catastrophic or existential risks to society,” including scenarios involving blackmail, resistance to shutdown, and unexpected behavior. Risks also include attackers hiding malicious instructions in data accessible by AI agents, potentially leading to data theft or unauthorized transactions.

Anthropic emphasizes that developing highly advanced models and expanding use cases could further increase the risk of harm. While the company is implementing safeguards like pre-release testing and red-team assessments, it acknowledges these efforts may not be sufficient. Furthermore, Anthropic may delay training or restrict model releases if safety measures are inadequate, a decision that could put it at a disadvantage against competitors. This cautious approach contrasts with the rapid deployment strategies seen elsewhere in the tech industry.

The company has also outlined an unusual governance structure featuring five different classes of stock, compared to Google’s three. The seven co-founders will retain 50.1% of the voting power on shareholder matters, while a separate Long-Term Benefit Trust will appoint a majority of the board. This structure indicates Anthropic’s intent to secure Wall Street funding while maintaining control over its core mission of safe AI development, even if it means foregoing lucrative business opportunities. Notably, Ben Mann, who co-leads the Labs team and is often referred to as a co-founder, is not part of the Founder LLC that will control the company post-listing.

In terms of customer engagement, Anthropic’s S-1 reveals that over half of its $100,000-plus customers now utilize both the Claude Platform and a first-party app, indicating a successful cross-selling strategy. Recent integrations include Invesco, CAIS, Conquest, and FINNY bringing their model portfolios and tools into Claude, while Dayforce integrated its HR, payroll, and scheduling data, allowing customers to query and act on workforce information using the AI.