Anthropic moved closer to a potentially record-setting IPO when it published its more than 300-page S-1, offering Wall Street its fullest look yet at a pure-play frontier AI lab. The headline numbers were hard to ignore: $16.29 billion in first-half 2026 revenue, an implied 85% gross margin and cash reserves surpassed only by Google and Berkshire Hathaway among non-bank U.S. companies.¹

The filing also set out the company’s preferred framing. As a public benefit corporation, Anthropic says investors should see a stake as more than a transaction—an opportunity to help it steer advanced AI safely toward “an age of miracles.”² Its founders call that an “anti-sell” strategy: radical honesty meant to draw shareholders willing to accept short-term sacrifice for a longer-term mission.²

But the same document made the price of that vision unmistakable. Anthropic has locked itself into at least roughly $850 billion in compute and infrastructure commitments over the next two decades, a figure that could approach $1 trillion with pending deals. PitchBook analyst Harrison Rolfes’s conclusion was blunt: “Growth is not optional for this company. The contracts assume it.”³

That pressure sharpened after the June figures. Revenue jumped from $4.7 billion in the first quarter to $11.6 billion in the second, then reportedly reached $16.8 billion in the third—still remarkable, but a slower sequential pace.⁴ Analysts say the unanswered question is whether each additional dollar of revenue will remain cheap enough to generate as compute demand expands.⁵

The risk is not solely financial. Anthropic warns that unintended model behavior could create liabilities beyond what insurers can cover, while specialists say cyber policies may address data breaches but leave bodily injury and physical damage as a “gaping hole.”⁶

Its governance adds another wrinkle: seven co-founders retain 50.1% voting power through a special Class F share, leaving public investors to decide how much control they are prepared to surrender.⁵ Inside the company, the tension is equally stark. Amid a possible $1.5 trillion-to-$2 trillion valuation, candidates are reportedly asked how they would react if safety concerns forced Anthropic to abandon its ambitions and its stock fell to zero.⁷

In a fragile IPO market and amid wider doubts over AI infrastructure spending, Anthropic is effectively asking investors to finance both its commercial sprint and its safety brake. Whether those promises can coexist will determine the value of the offering.¹