Anthropic, the company behind the Claude chatbot, has told prospective investors that the very technology it sells could pose a threat to humanity. The warning appears in its long-awaited initial public offering prospectus, which was shared with a select group of partners, according to people familiar with the filing cited by the Financial Times.
In the S-1 document filed with the US Securities and Exchange Commission, Anthropic devotes almost a third of its pages to risk factors. These include the possibility that increasingly advanced AI models could manipulate, blackmail, or behave unpredictably. The company's own research has shown autonomous models sabotaging code, assisting fraud, and manipulating information in controlled tests.
Such warnings complicate the task for investors trying to value a company whose listing, expected on the Nasdaq this autumn, is set to be the most highly valued IPO ever. Anthropic has not publicly commented on the reports.
Financial risks and customer concentration
The filing also flags significant financial risks. Close to a quarter of Anthropic's revenue last year came from just two customers, and many of its largest clients are not locked into long-term contracts. Revenue jumped twelvefold to almost $4.6 billion (€4bn) in 2025, but operating expenses reached $12.65 billion (€11.1bn), more than half of it on computing and infrastructure, leaving an operating loss of more than $8 billion (€7bn).
Its net loss of about $42 billion (€36.9bn) included a roughly $34 billion (€29.8bn) accounting charge linked to financing that could eventually convert into shares, rather than money spent running the business. Anthropic plans to spend $518 billion (€455.3bn) on cloud, computing, and infrastructure obligations in the coming years. It had $20.28 billion (€17.8bn) in cash and short-term investments at the end of 2025.
Revenue reached $11.5 billion (€10.1bn) in the second quarter of 2026, and the company is on course for a second straight quarter of operating profit, albeit on an adjusted basis. Backers expect Anthropic to list at a valuation above $2 trillion (€1.73tn), more than double the $965 billion (€848bn) set in its funding round in May. As Euronews reported earlier this month, at their peak, pre-IPO contracts on the crypto platform Hyperliquid were already implying a valuation of roughly $2.36 trillion (€2.05tn).
Safety fears reach the White House
The prospectus lands amid a heated debate over AI risks. Current and former Anthropic staff have warned that runaway AI could end humanity within a decade, and CEO Dario Amodei told the UN Security Council last week that AI was "the most important global security issue facing the world today."
Amodei has also led calls this month for "pacing the frontier" of AI development, backed by OpenAI's Sam Altman and Elon Musk. Yet, Anthropic released its Opus 5.5 model last week to counter OpenAI's momentum since the launch of GPT-6 Astra. OpenAI revealed last week that its tools had hacked "dozens" of external sites, including government ones, and on Monday scrapped plans to release its newest model next month, citing safety concerns.
US President Donald Trump rejected Amodei's call to slow down, saying: "I'm for let's go and let's win," candidly describing the AI race as a competition between the US and China that America is leading and cannot afford to lose. Amodei was left out of Thursday's White House state dinner with Chinese President Xi Jinping, attended by Musk, Altman, Nvidia's Jensen Huang, and Apple's Tim Cook among other executives. Trump then hosted him for a private dinner on Sunday night, their first one-on-one meeting, though the White House released no details about it.
The two sides have clashed since February, when Anthropic refused to let its models be used for fully autonomous weapons or domestic mass surveillance, which led the Pentagon to classify the company as a supply chain risk—a decision a federal appeals court upheld last week.
For European observers, the debate resonates with Pope Leo XIV's warnings that AI must serve humanity, not deepen injustice. The Vatican's concerns echo the broader European emphasis on ethical AI, as the EU continues to shape its AI Act. The outcome of Anthropic's IPO and the safety debate will have global implications, including for European businesses and regulators watching closely.
