OpenAI sent investors a memo claiming it has 1.9 gigawatts of computing capacity versus Anthropic's 1.4 gigawatts

Two AI rivals are fighting over who has more computing muscle as they prepare to sell shares to the public. OpenAI told its investors this week that it has more data center capacity than Anthropic, while Anthropic fired back with news about major deals and fast-growing sales numbers.

OpenAI shared a document with investors saying it currently has 1.9 gigawatts of computing capacity for 2025, which is three times what it had last year. The company said Anthropic has 1.4 gigawatts. OpenAI expects to reach somewhere in the low-double-digit gigawatts within twelve months and hit 30 gigawatts by 2030. The company predicted Anthropic would max out between seven and eight gigawatts before 2027 ends.

OpenAI’s document, which CNBC wrote about, argued that bigger infrastructure spending creates a cycle where better technology leads to lower costs, which then pays for improvements that bring in more customers.

The company also pointed to comments from Anthropic’s chief executive Dario Amodei, who last year said some competitors were “YOLO-ing” while his company took a careful approach. OpenAI suggested that caution was a mistake.

Anthropic responded by pointing reporters to statements from its chief financial officer, Krishna Rao, about a deal announced with Google and Broadcom earlier this week. That partnership will give Anthropic roughly 3.5 gigawatts of computing power starting in 2027. Rao said the company is making its biggest computing commitment yet to keep up with growth.

CoreWeave agreement adds more capacity

On Friday, Anthropic announced another deal with CoreWeave, a cloud infrastructure company, to get more computing capacity later this year. CoreWeave’s stock went up more than 5% before markets opened. The company has signed similar deals recently, including an $11.9 billion agreement with OpenAI last year, a $6.3 billion order with Nvidia in September, and a $21 billion deal with Meta on Thursday.

Anthropic also shared that its revenue has jumped to $30 billion per year, up from around $9 billion at the end of 2025. The company said more than 1,000 businesses now spend over $1 million each year on its services, double the 500 it reported in February. The company expects to reach positive cash flow by 2027.

Meanwhile, Anthropic has gained ground in business sales, as reported by Cryptopolitan previously. Its share of enterprise spending on artificial intelligence in the United States climbed to 40%, while OpenAI’s dropped from 50% to 27% during the same time. OpenAI has responded by focusing more on business customers and coding tools.

OpenAI has bigger plans for spending

OpenAI wants to put about $600 billion into chips and data centers through 2030, partly funded by a $122 billion fundraise. The company expects to lose $14 billion in 2026 and won’t break even until 2030.

Anthropic committed $50 billion to building computing infrastructure in the United States last November. The company runs its Claude models on different types of chips from Amazon, Google, and Nvidia, making it available on all three major cloud platforms.

The timing of OpenAI’s memo may raise questions as both companies prepare to go public. While OpenAI highlighted its larger infrastructure footprint, Anthropic is bringing in more money and spending far less to do it. Public market investors typically favor companies that can show a clear path to making a profit.

Anthropic projects it will reach that milestone by 2027, three years ahead of OpenAI’s 2030 target. The memo arrived just as Anthropic reported rapid growth in business customers, suggesting the competition between the two companies is heating up ahead of their stock market debuts.

The smartest crypto minds already read our newsletter. Want in? Join them.