Altman and Huang Seal $100 Billion OpenAI–Nvidia Pact After Intense Negotiations

Sam Altman and Jensen Huang at Nvidia’s Silicon Valley headquarters.

Nvidia and OpenAI have finalized a landmark $100 billion agreement that reshapes the global artificial intelligence industry. The deal was hammered out in a series of late-night meetings between OpenAI CEO Sam Altman and Nvidia CEO Jensen Huang, culminating just hours before Altman unveiled OpenAI’s next major infrastructure expansion in Texas.

A Race Against the Clock

President Trump meeting Sam Altman and Jensen Huang during UK visit.

According to people close to the discussions, the final terms were agreed at the last minute, following weeks of private conversations across London, Washington and San Francisco. The breakthrough coincided with President Donald Trump’s recent U.K. visit, where both leaders were able to brief the White House before returning to California for the public announcement.

Nvidia, OpenAI Forge $10B AI Supercomputing Pact

The pact cements Nvidia’s role as the most powerful hardware provider in the AI era, while giving OpenAI the resources to accelerate its 10-gigawatt supercomputing buildout. Nvidia will inject funding in stages—beginning with a $10 billion tranche—and supply its advanced GPUs to fuel OpenAI’s data centers.

Huang, whose company is now worth nearly $4.5 trillion, called the arrangement “monumental in size.” Altman, leading a startup valued at around $500 billion, described the partnership as critical for overcoming what he called “an unprecedented infrastructure challenge.”

Balancing Partnerships and Rivalries

The move arrives at a delicate moment in OpenAI’s ecosystem. Microsoft, OpenAI’s largest shareholder and primary cloud partner, was only notified a day before the deal was signed—months after losing its exclusive role as compute provider.

At the same time, OpenAI is committed to other multibillion-dollar partnerships, including a $300 billion cloud capacity agreement with Oracle beginning in 2027. These efforts are being consolidated under “Stargate,” OpenAI’s global infrastructure program backed by Oracle, SoftBank, and now Nvidia.

Financing the Expansion

The $100 billion pledge is just one piece of the financing puzzle. OpenAI is preparing to take on significant debt to cover the remainder of the infrastructure rollout, executives said, while seeking to avoid diluting equity. The staged structure ensures Nvidia’s investments grow alongside OpenAI’s valuation as new facilities come online.

A Long Relationship Comes Full Circle

Nvidia DGX supercomputer delivered to OpenAI office in 2016.

The collaboration is years in the making. In 2016, when OpenAI was still a small research lab, Huang personally delivered Nvidia’s first DGX supercomputer to its San Francisco office. Nearly a decade later, the two companies are at the center of the generative AI boom, driving global competition for compute power.

OpenAI's Data Center Push

The first new data centers are expected to launch in the second half of next year, though OpenAI and Nvidia declined to specify locations. More than 700 proposals are being reviewed across North America, with energy supply and permitting timelines among the deciding factors.

Executives have hinted that OpenAI may ultimately evolve into a first-party cloud provider, competing with Amazon Web Services, Google Cloud and Azure. Such a shift could arrive within two years, once OpenAI secures enough compute to meet its own demand.

Altman, speaking alongside Huang at Nvidia’s Silicon Valley headquarters, promised more announcements soon:
“Expect a lot from us in the coming months. The AI revolution is only just beginning.”

---