OpenAI's Stargate data centers are taking longer and costing more than competitors' facilities, according to a Distilled report by Michael Thomas. The infrastructure lag threatens OpenAI's ability to scale its AI models as fast as rivals like Google and Anthropic.
Key facts
- OpenAI raised over $40 billion in total funding
- Stargate data centers cost more than competitors
- Google invested $11B/year in SpaceX compute
- OpenAI competes with Google and Anthropic
- Codex reached 5 million weekly users in June 2026
OpenAI has raised over $40 billion in total funding, but its Stargate data center buildout faces cost overruns and extended timelines compared to competitors. According to Distilled | Michael Thomas, the Stargate projects are underperforming relative to industry benchmarks, though the report does not disclose specific cost or timeline figures.
Google, by contrast, has committed $11 billion per year to SpaceX for compute at xAI data centers, as reported earlier this month. Google Cloud also provides infrastructure for Anthropic, which has received $14 billion in investment from Google. OpenAI's infrastructure challenges come as it competes with both Google and Anthropic for AI market share.
The Infrastructure Race
OpenAI's Stargate program was announced as a flagship data center initiative to support its next-generation models, including GPT-5 and beyond. However, the Distilled report indicates that construction delays and cost overruns have made Stargate less competitive than similar projects from Google and other cloud providers. The company has not publicly disclosed the total investment or completion targets for Stargate.
This contrasts with Google's aggressive infrastructure spending: the company invested $11 billion annually in SpaceX compute alone, and its own data center buildout benefits from decades of operational expertise. Anthropic, meanwhile, leverages Google Cloud and has its own dedicated compute capacity.
Implications for Model Deployment
If Stargate timelines slip, OpenAI may face bottlenecks in training and deploying its largest models. The company's recent moves—including acquiring cloud startup Ona and considering API price cuts—suggest it is seeking alternative paths to scale. OpenAI's Codex product reached 5 million weekly users in June 2026, up 400% from the start of the year, adding pressure to deliver reliable inference infrastructure.
The cost overruns also raise questions about OpenAI's capital efficiency. With $40 billion raised, investors will scrutinize whether that capital is translating into competitive infrastructure or being consumed by delays.
What to watch
Watch for OpenAI's next infrastructure update—either a revised Stargate timeline or a new capacity deal with a cloud provider. Also track whether Google or Anthropic announce additional compute contracts in Q3 2026, which would signal a widening infrastructure gap.
Source: news.google.com
