OpenAI and Anthropic are competing in powerful data centers before investors' eyes

Large AI startups prepare for IPO: SpaceX (xAI), OpenAI and Anthropic

This year, the three biggest players in artificial intelligence—SpaceX (through its subsidiary xAI), OpenAI, and Anthropic—are considering going public. Although exact dates have not yet been announced, competition for attracting investment has already begun.

Competition for computing power
Recently, OpenAI and Anthropic staged a “victory” in the race to build their own infrastructure. According to CNBC and Bloomberg, OpenAI sent investors a memo comparing its data‑center plans with Anthropic’s ambitions.

- OpenAI aims to have access to 30 GW of compute by 2030.
- Anthropic, in turn, plans to reach only 7–8 GW by the end of 2027—“less aggressively,” according to OpenAI representatives.

In 2025, OpenAI already owned 1.9 GW of capacity, three times higher than the previous year’s figure. Anthropic finished last year with 1.4 GW of available resources.

Why expanding infrastructure matters
OpenAI explains the need to scale its computing network for several reasons:

1. Model optimization—more powerful centers allow refinement of AI models.
2. Lower token servicing costs—savings on each request increase margin.
3. Revenue growth—new models unlock additional income streams.
4. Company mission—to provide free AI tools to millions of users, which requires large-scale infrastructure.

By 2030, OpenAI plans to invest about $600 billion in developing its centers.

Anthropic’s stance on investment
OpenAI stated that Anthropic follows more conservative plans not out of caution but because it “underestimates the demand potential” for their services. This calls into question Anthropic’s growth strategy and increases pressure on competitors.

Bottom line: The three leading AI startups are already actively vying for investors and resources, with OpenAI clearly aiming to outpace Anthropic in the speed of computing infrastructure development. In the coming years, their strategies could significantly shape the artificial intelligence market.