Anthropic prepares for IPO with massive revenue growth and high compute costs

The path to a public listing

Anthropic has submitted its official registration paperwork to the U.S. Securities and Exchange Commission to seek a spot on the Nasdaq. The company plans to trade under the ticker symbol ANT. This move represents a major step as the AI firm moves from a private startup to a publicly traded entity.

The paperwork provides a detailed look at the company's financial health as it prepares for this transition. It highlights a period of intense growth and significant investment in the underlying technology that powers its Claude family of models. The company is navigating a complex landscape of high demand and high costs.

CFO Krishna Rao, who joined the company in 2024, is tasked with managing this transition. He must demonstrate that the company can turn its rapid revenue growth into a sustainable business model. The filing serves as a primary document for investors to assess the company's viability and long-term strategy.

Explosive revenue and heavy losses

The financial data in the filing shows a dramatic increase in scale for the AI company. During the first half of 2026, the firm brought in $16.29 billion, a sharp increase from the $1.16 billion reported during the same period a year earlier. This growth is driven by a rapidly expanding enterprise business.

Despite this growth, the company reported a net loss of $80.75 billion. A significant portion of this figure, specifically a $72.39 billion accounting charge, is tied to changes in the fair value of financing instruments. This type of charge can impact the bottom line without necessarily reflecting day-to-day operational performance.

Operating losses did narrow during the period, falling to $693 million from $2.14 billion. While the net loss is high, the narrowing operating loss suggests some progress in managing the core costs of the business. However, the company still faces a significant hurdle in reaching consistent profitability.

The cost of AI infrastructure

A primary challenge for Anthropic is the soaring cost of the computing power required to train and run its models. Spending on infrastructure hit $8.95 billion in the first half of 2026, a substantial jump from $2.33 billion in the previous year. The company relies on a mix of hardware from Amazon, Google, and Nvidia to power its operations.

CFO Krishna Rao has noted the importance of these costs, stating that the compute they procure is the "lifeblood of our business." He currently spends between 30% and 40% of his time focusing on compute infrastructure to ensure the company has the capacity it needs to grow.

Anthropic reports that it currently earns more than $2 for every $1 it spends on hardware costs. However, the company warns that this ratio could change if it is forced to lower its token prices to remain competitive. This risk highlights the delicate balance between maintaining high-quality AI and keeping costs low enough to generate profit.

Enterprise growth and market risks

The company's enterprise segment is showing strong momentum as of September 30. Over 11,900 clients contributed at least $100,000 in total earnings for the firm over the trailing 12 months. This includes a core group of approximately 250 customers who generated at least $10 million each.

While the growth is impressive, analysts point out several risks that could affect the company's trajectory. These include compute constraints, limited influence for shareholders, and potential conflicts between the company's public-benefit mandate and its financial goals. Investors will be looking closely at the growth trajectory into 2027 and 2028.

Anthropic also warns that falling behind competitors in developing advanced models could weaken demand. If the company cannot keep pace with the rapid evolution of the field, it could face a situation where infrastructure costs remain high while revenue plateaus. The upcoming IPO will be a test of whether the company can manage these pressures while scaling its operations.