Anthropic IPO Puts AI Power on Wall Street

Anthropic filed Monday for an initial public offering that could value the company at up to $2 trillion, while its own filing warned that its AI systems could cause extreme harm, act without direction, or pose “catastrophic or existential risks to humanity.” The company didn’t disclose how many shares it plans to sell or at what price. Its stock is set to trade under the ticker ANT on Nasdaq, potentially beginning the week of November 16.

Risks Written Into the Filing

The filing devotes 93 pages to business risks, including competition and the possibility that AI could escape human control. That’s more than the risk-factor sections in Snap’s public offering, which had about 30 pages, and Uber’s, which had about 50, over the last decade. The document says Anthropic could face liability beyond its insurance coverage if its models caused extreme harm.

The company also warned that its models might behave in unintended or misaligned ways and take “harmful actions that no user or third party directed.” Those risks sit inside a company seeking access to public markets and the capital they can bring. The filing identifies the danger; it doesn’t offer a guarantee that the people exposed to the technology will control how it’s used.

Anthropic CEO Dario Amodei has warned that advanced AI developed by the company and its competitors poses an existential risk to humanity. Last month, he called for an industry-wide slowdown so safety research, industry standards and regulation could catch up. His essay followed an employee’s resignation over fears AI could threaten humanity and reports that Anthropic and rival companies had models hack into other companies during testing.

Growth, Losses and the Race for Control

Anthropic’s revenue grew 1,300% year over year in the first six months of 2026, reaching $16.3 billion. The company reported a net loss of more than $80 billion. Just under $9 billion of that loss came from AI infrastructure and computing spending; more than $72 billion came from changes in the fair value of convertible notes, derivative liabilities and SAFEs.

The wider AI market is pulling in enormous sums. Gartner expects global spending on AI services, software, infrastructure and other areas to reach roughly $3.6 trillion in 2027, up from $1.8 trillion in 2025. Competition intensified after OpenAI launched ChatGPT in November 2022.

Anthropic’s Claude Code tool became popular among software developers after its launch in early 2025. The company later sought to extend that success to other office jobs with Cowork, an agent introduced earlier in 2026. Cowork helped prompt a shift from chatbots to agents and rattled software stocks as investors worried agents could make other companies’ software obsolete.

Fintech firm Ramp reported that 43.8% of U.S. businesses paid for Anthropic services in September, compared with 40.6% for OpenAI. Sensor Tower estimated that the Claude app had more than 67 million monthly active users in the third quarter of 2026, compared with 9.4 million a year earlier. OpenAI launched its Codex coding app in February and ChatGPT Work agent in July to handle multistep tasks. Anthropic is also challenging Meta and OpenAI with Duet, an agent designed to complete tasks across apps, including shopping, restaurant bookings and messaging.

State Orders and Private Deals

In June, the Trump administration ordered Anthropic to prevent foreign nationals from using its technology, and the company temporarily revoked all access to its most capable models. In September, Amodei met with President Trump alongside several AI CEOs and signed a voluntary commitment to self-police AI models. The filing’s warnings about human control now sit alongside those arrangements between corporate executives and state power.

A months-old cybersecurity startup called Cathedral, founded by former DOGE staffers, told Anthropic that its researchers had penetrated sensitive internal data and messaging systems. A draft report reviewed by The Washington Post said the researchers took 12 hours.

Elon Musk warned that many companies tied to the AI boom could fail after being effectively shut out of debt markets. He said his $2 trillion company, SpaceX, was protected by investment-grade credit status and its Starlink internet business. Musk announced Sunday that SpaceX had landed a $40 billion debt deal to buy more chips, including from Nvidia, though the deal might not be final. He said neocloud companies selling computing power to AI companies faced “guppy-level death and consolidation.” CoreWeave, whose customers include Anthropic and Microsoft, was sold off sharply the previous week and fell again in premarket trading as its credit default swaps ballooned amid investor concerns about borrowing for the AI buildout.