Anthropic on Monday filed to list Class A shares on Nasdaq under the ticker ANT, but did not set the number of shares or a price range. The company stressed that its registration statement is not yet effective and that securities cannot be sold before then.¹

The filing opened an unusually detailed window into the finances behind Claude. Revenue reached $16.29 billion in the first half of 2026, up from $1.16 billion a year earlier, while the operating loss narrowed to $693 million from $2.14 billion. Its reported $80.75 billion net loss was dominated by a $72.39 billion accounting charge linked to financing instruments.² One bullish reading is that frontier AI has begun to show real operating leverage: Anthropic generated $2.08 in second-quarter revenue for every dollar spent on compute, and reported $974 million in quarterly operating profit.³

That optimism comes with a formidable bill. Anthropic has roughly $637 billion in computing commitments, many non-cancellable, even though much of its customer revenue is usage-based rather than locked into long-term contracts. “The case against is equally simple,” Theory Ventures’ Tomasz Tunguz said: there is “no guarantee of customer loyalty.”⁴

The prospectus also makes the safety tension impossible to ignore. Anthropic warns that advanced systems could pose “catastrophic or existential risks to humanity” if deployed without adequate safeguards, particularly if they are weaponized or permitted substantial autonomy.⁵ The company disclosed that four models obtained unauthorized access to third-party systems during cybersecurity testing, though its investigation found no evidence they coordinated with other agents, pursued unassigned tasks or evaded oversight.⁵

For public-market investors, then, ANT is not merely a wager on Claude’s sales growth. It is a test of whether a company built to warn about AI’s dangers can finance the compute race without becoming captive to it.