What Anthropic's IPO filing says about revenue, net loss and investor valuations

Evercore has expressed optimism regarding Anthropic's S-1 filing, noting that the financial disclosures paint a favorable picture for cloud service providers that collaborate with the artificial intelligence company. Anthropic's revenue has exhibited substantial growth over the past three years, escalating from $41 million in 2023 to $386 million in 2024, $4.6 billion in 2025, and reaching an impressive $16.3 billion in the first half of 2026.

The S-1 filing indicates that this rapid expansion has led to a capacity constraint, with demand for computing resources surpassing what is currently available. In response, Anthropic has been compelled to seek additional compute power from its cloud partners, who were initially reluctant to provide it. The company's customer base is extensive, encompassing every Fortune 10 corporation and over half of the Fortune 500, along with $100,000-plus customers and 40 million smaller users.

This diverse clientele extends across both enterprise and consumer sectors. Anthropic's valuation is grounded in the global knowledge-worker market, estimating it to be worth over $40 trillion in 2025, which serves as the company's total addressable market for AI solutions. Recent updates to Anthropic's S-1 filing have unveiled a staggering 1,368% increase in revenue during the second quarter compared to the previous year, while compute and infrastructure expenses have surged by 300% over the same period.

Furthermore, the filing emphasizes Anthropic's substantial commitments of $200 billion in agreements related to NVIDIA-based compute infrastructure and partnerships with cloud service providers and chip manufacturers. However, Yorkville Ives has cautioned that Anthropic faces potential challenges stemming from compute capacity limitations and take-or-pay obligations that could impede the company's growth within the artificial intelligence market.

Additionally, Anthropic has integrated its Claude AI assistant with Google Workspace applications, enabling users to access the tool directly within Google Docs, Sheets, and Slides. Moreover, Anthropic's CEO, Dario Amodei, is set to deliver a briefing to the United Nations Security Council on AI-related risks, joining a conversation with OpenAI CEO, Sam Altman.

On a separate note, China's internet regulator is currently investigating claims that DeepSeek and Moonshot AI transmitted sensitive data to Anthropic's Claude models without proper authorization.