Nvidia and OpenAI signed a $100B deal to build data centers and buy OpenAI equity

Nvidia and OpenAI signed a $100 billion agreement on Monday that sounds more like a handshake tailored for algo-driven markets than an actual business necessity.

The setup is that Nvidia will sell chips to OpenAI for giant data centers. In return, it’ll spend $100 billion buying OpenAI’s unlisted stock over time, far more than the $72 billion OpenAI has raised in its ten-year history.

Both companies claim this move will help accelerate AI development, but the scale of the deal and the way it was announced made it feel more like financial performance art than a genuine requirement.

On paper, this arrangement locks Nvidia into supplying chips for AI models like OpenAI’s GPT-5, which are set to run on a new wave of infrastructure. But then comes the odd part; Nvidia doesn’t need to do this.

The company is already the top supplier of advanced AI chips, and its closest competitors are still struggling to catch up. Even OpenAI has been trying to develop its own chips, but for now, it’s still heavily dependent on Nvidia.

Nvidia invests $100B and builds 10GW data centers for OpenAI

The deal includes the construction of data centers powered by Nvidia chips with a minimum capacity of 10 gigawatts. These aren’t just for testing. They’ll train and deploy real AI models.

And while Nvidia will provide the hardware, it’s also going to be an investor, piling in $100 billion over time to increase its existing stake in OpenAI. This equity-based structure means OpenAI won’t owe anything back. If the value drops, or chip orders don’t follow through, Nvidia absorbs the loss.

Jensen Huang, CEO of Nvidia, isn’t betting the company. With about $100 billion in yearly free cash flow and a $4.5 trillion valuation, he can take a hit. But the real question is why do it at all? Monday’s announcement bumped Nvidia’s market cap by $180 billion. Not a bad day in the markets, but a small gain for a company that size.

Sam Altman’s side of the story is clearer. OpenAI is expected to generate $12 billion in revenue this year, but it can’t cover the massive capital expenses needed to stay ahead. For a company claiming to lead the AI race, raising money in chunks should be easy.

Still, the whole thing reeks of performance. OpenAI wants to look unstoppable in the race toward superintelligence. Its $500 billion valuation depends on that belief. Big data center plans create momentum. For Nvidia, being seen as the go-to AI chip supplier forces rivals to panic-buy before it’s too late.

Nvidia calms partners as competitors rethink chip strategy

Right after the news broke, Nvidia tried to calm the rest of its customer base. “Our investments will not change our focus or impact supply to our other customers,” the company said in a statement. “We will continue to make every customer a top priority, with or without any equity stake.”

That statement was clearly aimed at companies like Microsoft, Meta, Amazon, and Alphabet; all of whom Nvidia still heavily depends on for revenue. These firms are racing to build AI infrastructure too.

Many are developing their own chips to reduce reliance on Nvidia, but they’re still buying from Jensen’s team in the meantime. Now, those same companies are watching the OpenAI–Nvidia relationship closely.

If Nvidia is seen as giving OpenAI special access, competitors may push harder to go independent. The fear of falling down the priority list could push them to speed up internal chip development or look more seriously at Nvidia alternatives like AMD.

Your crypto news deserves attention - KEY Difference Wire puts you on 250 top sites