Anthropic files for IPO with $11.56 billion quarterly revenue

Anthropic filed for an initial public offering on Monday, positioning the five-year-old artificial intelligence company to pursue what could be the largest stock offering in history. The filing reveals that second-quarter revenue reached $11.56 billion, a 14-fold increase from $787 million a year earlier. The company reported $974 million in operating income for the same period, marking its first profitable quarter after five consecutive quarters of operating losses.

Investors are assessing the company's potential market cap in the trillions of dollars, placing it in the category of the world's most valuable firms. Anthropic would become the third company to debut with a valuation above $1 trillion, following Saudi Aramco and SpaceX, which listed in June. OpenAI, which has confidentially filed to go public, has stated its debut is unlikely before 2027.

The company's financial structure is deeply intertwined with major technology firms. Nearly half of Anthropic's revenue is generated through cloud divisions at Microsoft, Amazon, and Alphabet, all of which hold investments in the company. In connection with the IPO, Anthropic can issue up to $77 billion in notes to Amazon, Broadcom, and Google. Additionally, Microsoft and Nvidia have agreed to invest $15 billion to expand Anthropic's computing capacity. Chips from Nvidia, Advanced Micro Devices, Amazon, and Google power the company's development.

Despite strong revenue growth, Anthropic recorded over $80 billion in losses during the first half of the year. Jai Das, co-founder of Sapphire Ventures, noted that if the offering underperforms expectations, the situation could become unstable for the sector. He compared the current infrastructure spending to the telecommunications bubble of the late 1990s, stating that while some companies failed during that period, the infrastructure eventually found demand. Das believes there is no long-term lack of demand for AI services.

Cloud providers are increasing capital expenditures to meet this demand, while memory manufacturers face shortages prioritizing AI chips. Alphabet has committed to backstopping a data center project in Texas, and Broadcom is creating financing pools to help Anthropic and OpenAI deploy high-performance models. For public market investors, more than one-third of the S&P 500's value comes from tech companies with trillion-dollar valuations that are heavily investing in AI. Software giants like Adobe and Salesforce have also integrated with Anthropic's Claude assistant.

In private markets, Bessemer Venture Partners and Menlo Ventures identified Anthropic as their largest investment ever. Brad Bernstein, a managing partner at FTV Capital, warned that if Anthropic does not perform well, it will become a bellwether for the AI sector, potentially forcing a reset of valuations across venture-backed AI companies. The prospectus identifies OpenAI and Google as primary competitors but notes broader competition from open-weight and open-source models.

Open-weight models present a specific concern due to lower costs and customization flexibility. However, Michael Monaghan, portfolio manager at Founder ETFs, argued that it is too early to determine if these models will displace frontier models. He stated there will be room for both categories. Snowflake CEO Sridhar Ramaswamy said his company is experimenting with low-cost open-weight models from Chinese firms Moonshot AI, DeepSeek, and Zhipu to maximize internal efficiency. Despite this experimentation, Ramaswamy expressed support for Anthropic and OpenAI succeeding as public companies.

David Trainer, CEO of New Constructs, wrote that the AI boom relies on cheap credit being readily available. He warned that if AI liquidity dries up, there is a 20% downside risk in the S&P 500, concentrated in AI-related stocks. Rising interest rates are already making debt financing more expensive for companies seeking capital.