OpenAI's Stargate project - the $500 billion joint venture announced in January 2026 with SoftBank, Oracle, and other backers - is moving from press release into construction. The company published an update outlining how it's adding data center capacity across the United States to meet growing demand for its AI products.
The framing is ambitious: OpenAI describes this as infrastructure for the "Intelligence Age," its term for the period leading to artificial general intelligence (AGI) - systems capable of performing most cognitive tasks a human can do. The update doesn't announce a specific new facility or partnership, but documents progress on building the physical foundation for running and training increasingly large models.
The original Stargate commitment was $100 billion in the first phase, scaling to $500 billion over four years. Data centers at this scale typically take 18 to 36 months from groundbreaking to operational, which means what OpenAI is describing now will mostly come online in 2027 and beyond - relevant for the next generation of models, not the ones running ChatGPT today.
For daily users, more infrastructure capacity means fewer outages, faster response times, and the ability to deploy more capable models without the rate limits and slow rollouts that have frustrated power users. OpenAI's compute demand has consistently outpaced supply since GPT-4 launched; this buildout is their answer.
The competitive stakes are real. Google and Microsoft are making similar bets, each committing tens of billions annually to AI compute. The company that can run the most powerful models at the lowest cost per query will have a durable pricing advantage - and pricing is increasingly how these companies compete for enterprise contracts.