Anthropic is presenting investors with an increasingly attractive financial story ahead of its expected blockbuster IPO: explosive revenue growth, improving margins and positive adjusted operating income for a second consecutive quarter.
There is, however, a major catch.
The profitability measure excludes some of the expenses most fundamental to running a frontier artificial intelligence company – including the enormous cost of training new AI models and stock-based employee compensation. That distinction is drawing scrutiny as investors attempt to determine how profitable Anthropic’s rapidly growing business actually is.
Anthropic generated roughly $11.5 billion in second-quarter revenue, about 14 times the $787 million recorded a year earlier and up sharply from $4.73 billion in the first quarter.
The company has told investors it expects positive adjusted operating income for a second straight quarter, according to reporting cited by Reuters.
Anthropic’s Definition of Profit Matters
Adjusted operating income can be useful for understanding the performance of a company’s underlying operations, but it is not equivalent to bottom-line net income.
In Anthropic’s case, the difference is particularly important because developing frontier AI models requires enormous amounts of computing infrastructure.
The company’s adjusted figure excludes costs associated with training AI models as well as stock-based compensation, according to Futurism’s account of the company’s investor disclosures. Anthropic therefore remains deeply in the red under broader measures of profitability.
That makes statements about Anthropic being “profitable” highly dependent on which accounting measure is being discussed.
The company can point to a rapidly improving operating business and extremely strong revenue growth. Critics, however, argue that excluding model-training costs from a profitability metric is difficult to overlook when training increasingly capable models is central to Anthropic’s competitive position.
Training AI Models Is Extremely Expensive
Those costs are hardly marginal.
Anthropic has previously projected that training expenses alone could exceed $12 billion in 2026 and approach $23 billion the following year. Between 2026 and 2029, its projected model-training spending was expected to exceed $100 billion.
The company is simultaneously signing enormous infrastructure agreements to secure enough computing capacity for Claude.
In August, Anthropic agreed to spend about $45 billion over six years renting AI computing capacity from Nscale’s West Virginia data center, according to Reuters. The project is expected to use Nvidia’s Vera Rubin chips and provide Anthropic with 460 megawatts of capacity.
Anthropic also relies on infrastructure relationships with Amazon, Google, Microsoft and other providers as it races to expand capacity.
Those commitments illustrate why investors may focus on cash generation and conventional profitability alongside adjusted operating metrics.
Revenue Growth Is Still Extraordinary
None of this diminishes the scale of Anthropic’s growth.
The company’s annualized revenue run rate exceeded $65 billion by the end of July, up from roughly $9 billion at the end of 2025.
That trajectory has helped support expectations for a potentially historic public offering. Anthropic is discussing an IPO that could raise as much as $100 billion at a valuation around $2 trillion, with Nvidia considering an investment of up to $10 billion as an anchor investor.
That makes the debate over profitability more than an accounting argument.
At a potential $2 trillion valuation, public-market investors will have to decide how much weight to place on Anthropic’s extraordinary revenue expansion versus the enormous capital requirements needed to maintain its technological position.
Adjusted operating profitability suggests the commercial side of Claude is becoming increasingly powerful. But model training, infrastructure and employee equity are not peripheral to Anthropic’s business – they are part of what allows the company to compete at the frontier of AI.
As Anthropic approaches the public markets, investors will therefore be watching not simply whether the company can describe itself as profitable, but how much cash it ultimately needs to keep generating the growth that could justify one of the largest IPO valuations in history.
