Anthropic's landmark IPO prospectus, made public on Monday, lays out both the strength and the fragility of the artificial intelligence company's business. The S-1 is the registration statement a company files with the U.S. Securities and Exchange Commission before its shares can be offered to public investors, and it obligates the filer to lay out, in its own words, the material risks facing the business. The filing shows that customers are using Anthropic's AI services at extraordinary rates—but it also makes clear that the company is locked into hundreds of billions of dollars in computing commitments while many of its customers remain free to reduce their usage or terminate at nearly any time.

Anthropic built its fast-growing business largely by charging customers based on how much they use its services. As the prospectus made clear, however, the company is highly dependent on outside partners for the computing power that runs those services. The disparity between the agreements through which Anthropic buys computing power and the way it sells its AI services to customers means it could get squeezed if business conditions change.

Explosive Revenue Growth

The company's revenue is expanding at an exceptional pace. First-half revenue surged 1,300% to $16.28 billion, with roughly $13.1 billion of that—about 80%—coming from enterprise customers paying consumption-based fees rather than individual subscriptions.

The growth reflects booming demand for AI services, as businesses deploy advanced large language models such as Anthropic's Claude family of products to help with computer coding, accounting, and other tasks.

To fulfill that demand, Anthropic has signed massive infrastructure agreements with Amazon, Microsoft, Google, and other hyperscalers. Those cloud providers control the data centers and chips on which Anthropic's AI models run.

So far, the amount Anthropic pays these providers for compute capacity has grown at a more modest clip than its exploding sales figures. The company's cost of revenue for the tokens it serves customers—tokens being the small units of text that AI models process—was $2.4 billion in the six months ended June 2026, up 135% from the same period a year earlier. It also spent a hefty $9.7 billion on computing and infrastructure costs for research and development of future models.

Anthropic currently says it makes more than two dollars in revenue for every dollar it spends on computing infrastructure. But that equation could change if the company is forced to reduce token prices—a risk it explicitly warns about in its S-1 filing.

The Mismatch Between Spending and Revenue

Anthropic's infrastructure spending is committed well ahead of expected demand. The company has committed at least $475 billion in future cloud compute purchases with its major partners, including at least $389.9 billion pledged to Amazon, Google parent Alphabet, SpaceXAI, and Microsoft alone, alongside deals with Nscale and Lambda. Much of that money is owed regardless of how much usage its AI actually attracts. When deals to lease computing chips are factored in, its total commitments climb to as much as $750 billion.

The demand side of the business enjoys far less certainty. As the company laid out in the S-1, much of its revenue lacks long-term commitments, with contracts allowing customers to reduce usage or terminate with limited or no notice.

That mismatch between how Anthropic generates revenue and what it expects to spend creates a pressure-filled balancing act for the years ahead. Its growth has so far helped it keep pace with compute costs. The question now is whether customers will keep spending at a rate that allows the company to absorb these massive commitments.

"We must grow our customer base and expand revenue rapidly in order to generate the cash flows necessary to satisfy these commitments," Anthropic said in its S-1.

The company also acknowledged that hitting those growth targets could be a challenge. "We expect our revenue growth rate to decline over time and to fluctuate from period to period," it wrote.

Recognizing the potential mismatch, Anthropic has increasingly tried to move enterprise customers into noncancelable, multiyear deals with fixed minimum spends in exchange for discounted token rates. The company said that as of June 2026, it had about $10.8 billion in remaining performance obligations under these contracts—an accounting measure of revenue already under contract but not yet recognized. However, 75% of that revenue will be recognized over the next year, while Anthropic's data center spending commitments extend well beyond that time frame.

Anthropic also disclosed that it has significantly underestimated its compute requirements in the past, leaving it capacity constrained and impacting product development. The timing of compute coming online may not align with demand, it wrote, resulting in "periods of excess or underutilized compute and committed cash outflows that do not produce immediate revenue."

In the two weeks that ended Jan. 2 of this year, for example, Anthropic noted that customer spend declined by about 20% compared with the prior two weeks, although it later bounced back.

Anthropic typically contracts for compute 12 to 18 months in advance, and its expenditures rise as capacity becomes available. The risk is that even as revenue continues to grow, that growth could prove too slow to absorb the next wave of infrastructure costs. In its filing, Anthropic explicitly warns that significant quantities of new compute coming online will compress its margins.

For anyone tracking that balancing act going forward, the filing itself supplies the benchmarks: the ratio of revenue to compute spending, the growth of committed multiyear contracts, and the pace at which the hundreds of billions in cloud obligations come due are all figures the company has now placed on the public record.

"The numbers are growing at an unprecedented rate," Gil Luria, an analyst D.A. Davidson, told Fortune. "No company in history has ever grown this fast, this efficiently. They believe they can continue to grow, and they believe so based on their current success."

Luria added that despite any risk Anthropic carries in its customer contracts, the company has shown it is profitable and has ample cash at its disposal.

Significant Upfront Payments

The good news for Anthropic is that, so far, its plan has worked. Revenue has kept pace with compute expenses, and in its second quarter the company generated $974 million in operating income. It also held $112.4 billion in cash and cash equivalents as of late June.

Anthropic disclosed that its funding plans include operating cash flows, equity, debt, and partner financing. It warned, however, that its infrastructure build-out will increasingly require significant upfront payments before computing capacity comes online, threatening to eat into cash reserves even as financial performance improves.

The company is also not counting on advertising to boost revenue anytime soon, stating that it wants to avoid pressure to optimize for engagement. It cited its governance structure as a public benefit corporation, which obligates it to adhere to a public benefit mission alongside shareholder interests.

An Industry-Wide Balancing Act

Weighing AI demand against infrastructure spend has been a central issue across the industry as tech giants rush to build out data centers. The largest players have spent considerable effort catching up to demand while betting on a future in which enterprises and individuals continue to rely on AI services for a wide array of tasks. Amazon, Microsoft, Alphabet, and Meta are collectively expected to devote more than $700 billion to capital expenditures in 2026.

Many billions of that spending are being financed by debt, according to a report in June by England-based Exponential View. Oracle stands out in particular for its aggressive fundraising. Earlier this year, the company said it would need to raise between $45 billion and $50 billion in 2026 alone to fund its infrastructure expansion.

Other Risk Factors

Anthropic's delicate balance between revenue streams and fixed obligations is one of many risk factors highlighted in its extensive S-1. Others include reliance on a small number of clients: its top five customers accounted for roughly 25% of revenue in the first nine months of this year.

Anthropic also disclosed incidents in which its models gained unauthorized access to real-world systems during security testing, and estimated that 5% to 7% of its revenue this year came from customers later found to be violating its usage policies.

This article originally appeared on Fortune.com.