Anthropic filed paperwork Monday for a stock market listing that could value the AI company at more than $2 trillion, a debut that would surpass SpaceX's roughly $1.8 trillion valuation this summer as the largest in history — and the company used the filing to warn investors that the technology it sells could, in the wrong hands, threaten humanity itself.
The Claude chatbot developer will list on the Nasdaq under the ticker ANT, becoming the first fully AI-focused company to go public and a test case for how Wall Street prices the industry's enormous spending against its enormous risks. Anthropic has not yet disclosed how many shares it will sell, their price range, or a specific listing date.
Billions in losses, more cash than Meta and Google combined
The 335-page filing shows Anthropic lost $80.7 billion in the first half of 2026 — more than double what it lost across all of 2025 — against revenue of $16.3 billion over the same period. The company nonetheless holds $131.6 billion in cash, nearly all of it classified as working capital, more than Meta and Google hold combined. "In general, we will aim to be bold but sane, and to win trust with a track record of successful bets," Anthropic told prospective investors, acknowledging its heavy forward investments "may seem irresponsibly large at first glance."
Troy Hooper, co-head of equity capital markets at Mergermarket, said Anthropic's revenue growth is "exceptional," but that investors are "less forgiving" of lossmaking tech companies than they were months ago, adding that heavy accounting charges tied to certain investments leave the company's "underlying profitability" unclear.
A warning about its own technology
Across roughly 90 pages of risk disclosures, Anthropic said that "sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponised, or allowed to operate with a high degree of autonomy." The filing disclosed that a hacking group Anthropic believes was state-sponsored had, within the past year, manipulated Claude to breach a number of technology companies, financial institutions, chemical manufacturers and government agencies.
The company also flagged its dependence on a small cluster of cloud providers, Taiwanese chipmakers already stretched by demand across the industry, and data centers concentrated in the US, along with the risk of future "regulatory action arising from harms associated with our AI models," including use by children or people at risk of suicide. Anthropic's main rival, OpenAI, is expected to pursue its own listing early next year.
The filing landed the same day the owners of Elvis Presley's music catalog sued Anthropic for allegedly training Claude on 274 of his songs without a license — one of several copyright suits the company now faces, following a $1.5 billion settlement last year with authors who said their work was used without permission, the largest publicly reported copyright recovery in history.
