OpenAI scales up data-center ambitions with $500B Stargate push

OpenAI is accelerating its Stargate AI infrastructure initiative, a multi-year, multi-hundred-billion-dollar bet that the world will need vastly more compute than currently exists.

What Stargate actually is

The initiative, announced in January 2025, targets $500 billion in AI infrastructure investment over four years.

The latest milestone came in June 2026, when OpenAI held a groundbreaking ceremony for a $16 billion Stargate campus in Saline Township, Michigan.

Partners include Oracle, Microsoft, Nvidia, and AMD.

OpenAI paused its UK Stargate project, citing energy costs and regulatory concerns. The pause affects plans that had included roughly 31,000 GPUs. Separately, the company stepped back from expansion discussions in Texas.

The money behind the machine

In April 2026, the company closed a funding round that raised $122 billion, pushing its valuation to $852 billion.

OpenAI reported $2 billion in monthly revenue as of April 2026, alongside 900 million weekly active ChatGPT users.

Where crypto fits in this story

Bitcoin mining firms are pivoting toward AI data centers. The economics make sense: Bitcoin mining requires massive power infrastructure, high-density computing facilities, and sophisticated cooling systems. AI data centers require exactly the same things.

Core Scientific is one concrete example. The company is planning to lease six data centers to CoreWeave for a projected $10 billion in revenue across a 12-year term.

What investors should be watching

For investors already holding positions in Nvidia or AMD, the Stargate partnerships provide a fairly direct line of sight to ongoing demand. Both companies are embedded in the initiative and positioned to supply the GPU capacity that makes these campuses functional.

Microsoft’s involvement is worth watching separately. The company’s existing cloud partnership with OpenAI already gives it unique exposure to ChatGPT’s 900 million weekly users.

Energy availability and regulatory friction are now genuine constraints on how fast the AI infrastructure build-out can proceed, as demonstrated by the pauses in the UK and Texas.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.