The amended S-1 that will anchor Anthropic's public debut is, on its surface, a growth story: $17 billion in quarterly revenue and a year-over-year expansion that makes even the best-performing software companies look sedate. The number that matters, however, is not on the income statement. It is a $895 billion purchase and lease commitment, almost entirely compute, a figure that sits 27 times larger than the company's nine-month revenue.
The commitment structure is the story that a headline revenue number cannot tell. Nearly all of the $895 billion is tied to hardware and capacity that the company did not build and does not control. Roughly $70 billion of it comes due within a year, which converts a growth narrative into a liquidity test. Public investors who underwrite a listed company are underwriting those obligations, not just the run rate of sales.
The operating profile reinforces the tension. Reported operating income turned negative in the most recent quarter, although a large share of the net loss is noncash, driven by convertible instruments gaining value rather than by cash burn. Adjusted operating margin, at roughly 15 percent, is down from a year earlier, and the company attributes the squeeze to more expensive short-term compute and a donation-matching charge that will not repeat at the same scale.
There is a second-order market already forming around the listing itself. Prediction platforms have opened boosted-trading contests tied to the offering, and retail chatter has fixed on a single question: whether the ticker trades on the growth print or on the commitment stack. That split, the gap between the two readings of the same document, is the entire risk of the listing distilled into one decision.
For the broader AI industry, the filing sets a template. When a frontier lab lists with its balance sheet, its infrastructure obligations become part of the public record and, eventually, of the credit markets that price them. The roadshow argument is no longer only about safety; it is about whether a stated commitment of this size is a feature of the business model or the first thing public markets learn to discount.
Image source: i.ibb.co