Anthropic flags catastrophic technology risks as it prepares public debut

Artificial intelligence developer Anthropic has warned prospective investors that the technology it builds could present catastrophic or existential dangers to humanity, according to an initial public offering filing submitted on Monday. The cautionary disclosure comes as the company prepares a stock market debut that could raise as much as $100 billion.

The San Francisco firm, known for creating the Claude chatbot, intends to trade on the Nasdaq exchange under the ticker symbol ANT. If it reaches its target, the listing would overtake the $86 billion raised by SpaceX in June to rank as the largest initial public offering on record. Market estimates suggest Anthropic could achieve an overall valuation between $1.5 trillion and $2 trillion, up from a private valuation of $965 billion recorded earlier this year.

In the regulatory submission, the company outlined several major hazards tied to increasingly capable systems. These included risks that future models could assist in cyberattacks, contribute to the creation of biological weapons, empower mass surveillance, or operate beyond human supervision. The filing also cautioned that advanced models might display unpredictable behaviors, such as obscuring information, resisting attempts to shut them down, or carrying out harmful tasks, though Anthropic presented these as potential perils rather than certain outcomes.

Anthropic was established in 2021 by former OpenAI executives, including siblings Dario and Daniela Amodei, and has long emphasized technical safeguards as a primary focus. Dario Amodei, the chief executive, previously argued in a September essay that the tech sector should intentionally decelerate the creation of frontier models. He stated that buying even an additional year or two before reaching critical capability milestones could significantly lower catastrophic risks by giving researchers and policymakers more time to align artificial systems with human safety standards.

The company's push for new public capital coincides with surging expenses. Financial data in the filing showed that Anthropic posted a net loss of $80.7 billion during the initial six months of 2026, widening steeply from a $13 billion loss over the same period a year earlier. Operating costs have mounted rapidly due to enormous investments in computing power, custom microchips, specialized data centers, and advanced research.

The company indicated that funds raised from the public debut would help support infrastructure growth, pursue potential acquisitions, and maintain its competitive stance against rivals such as OpenAI and Google. The public listing would also stand to create massive paper wealth for Anthropic's seven founders, who together control nearly 500 million shares.

Beyond technological dangers, the documentation warned prospective buyers of multiple commercial headwinds, including strict regulatory scrutiny, stiff market competition, high employee turnover risks, and constraints on accessing necessary computing hardware. Anthropic noted that meeting heightened safety and oversight requirements may ultimately increase operational expenses and affect the pace of commercial rollout.