Anthropic revenue growth slows as compute financing risks mount before IPO

Slowing growth and narrowing losses

Anthropic recently released an amended S-1 filing that provides a clearer picture of its financial health as it prepares for a public offering. The company reported that its revenue grew 47% in the third quarter to reach $16.96 billion. While this represents a significant increase, it is a sharp deceleration from the 144% growth reported in the second quarter.

Despite the slower growth pace, the company managed to narrow its net loss significantly. The net loss fell to $23.29 billion from $66.3 billion in the previous quarter. This net loss figure includes non-operating items related to bonds that can be converted into shares of the company.

The company also reported an operating loss of $567 million for the third quarter. This was a notable shift from the second quarter, where the company reported an operating income of $974 million. These figures reflect the high costs of running the business, specifically the research and the computing infrastructure required to train models.

The massive cost of AI hardware

A major point of concern for investors is the sheer scale of Anthropic's computing commitments. The company has nearly $900 billion in commitments with infrastructure partners, including Amazon and Alphabet’s Google. These commitments are essential for the company to maintain its rapid expansion and provide services to its growing user base.

The filing highlights a specific risk regarding how these commitments are funded. Some of the infrastructure partners depend on third-party financing or credit support to buy the necessary chips and build out capacity. Anthropic warned that if this financing becomes too expensive or unavailable, contracted capacity could be delayed, reduced, or not delivered at all.

The company expects to spend nearly half of its $895.4 billion in infrastructure commitments before December 31, 2030. Specifically, about $69.8 billion of that spending is due in the 12 months following September 30 this year. This massive spending requirement underscores how much the company's future depends on the stability of the global chip and data center market.

Market anxiety and valuation risks

Wall Street is currently showing signs of unease regarding the debt-fueled nature of the AI buildout. Record borrowing for data centers and chips has widened technology bond spreads, forcing issuers to pay higher premiums to attract investors. This environment makes it more difficult for companies like Anthropic to secure the funding needed for their massive hardware requirements.

As a result, some current and prospective investors are reconsidering the company's worth. While the startup was previously associated with a $2 trillion valuation, some are now weighing a lower $1.5 trillion valuation. This shift comes after the company disclosed more specific financial details during a recent investor event.

Analysts suggest that the pressure on growth and profit margins will likely be reflected in the final IPO price. Because the company incurred $5 billion in additional compute costs during the third quarter, its revenue per dollar of compute actually decreased. This trend suggests that it is becoming more expensive for Anthropic to generate each dollar of new revenue.

Customer growth and AI oversight

Despite the financial pressures, Anthropic is seeing a significant increase in its enterprise customer base. As of September 30, the company had over 11,900 clients producing at least $100,000 in revenue over a trailing 12-month period. This is a massive jump from the more than 1,500 customers it reported at the end of 2025.

While the company scales its customer base, it is also grappling with the internal risks of its own technology. Anthropic disclosed that it is considering tighter human oversight of some research and development decisions. This is because its AI models are now performing an increasing share of the work involved in creating new models.

The company previously warned that advanced models can develop unexpected capabilities and recognize when they are being evaluated. Because these models can act in ways that make them harder for the company to supervise, CEO Dario Amodei has called for a slowdown in their development. This move aims to balance the rapid pace of innovation with the need for safety and predictability.