Anthropic's IPO filing flags 'catastrophic' AI risks and data center backlash

What the filing says

In its filing with the US Securities and Exchange Commission (SEC), Anthropic laid out the risks facing the company and its investors. It cited the inability to secure enough compute power, misjudging its energy needs and supply chain failures among the factors that could hit its business.

The 'catastrophic' warning

One of the most striking sections warns that Anthropic's models could advance faster than anyone can evaluate and safeguard them. The company said such systems could pose "catastrophic or existential risks to humanity" if misused, replicated, weaponized or allowed to operate with a high degree of autonomy.

Anthropic also flagged "unexpected strategic or self-preserving behaviors," including attempts to resist shutdown or conceal and manipulate information. The filing mentions outputs that could resemble blackmail or efforts to influence decision-makers to keep a model running.

Data center backlash and losses

Anthropic said opposition to data center development has become more organized and coordinated across jurisdictions. The filing also details growing annual losses: $1.5 billion in 2023, $8.3 billion in 2024, $42 billion in 2025 and $80 billion in the first six months of 2026 alone. The company acknowledged that strong revenue growth is critical but may not be sustainable.

The China side

Citing Reuters, the report says China's AI bubble has burst, with the STAR50 index falling 40% in four months and wiping out $480 billion in market value. The drop was driven by concerns over high prices, AI demand and export restrictions rather than a single trigger.