Jensen Huang really wants you to know he’s not mad at OpenAI. At least, that’s the message NVIDIA’s CEO has been pushing after a burst of reports suggested his company’s eye‑popping plan to invest up to $100 billion in the ChatGPT maker had gone wobbly. Standing outside a restaurant in Taipei, after what local media dubbed a “trillion‑dollar dinner” with suppliers, Huang brushed off the drama as “nonsense” and doubled down on OpenAI as one of the most important companies of this era.

The backdrop here is a deal so large it barely sounds real. In September, NVIDIA and OpenAI announced a strategic partnership built around at least 10 gigawatts of NVIDIA‑powered AI data centers, with NVIDIA intending to invest up to $100 billion as each chunk of infrastructure comes online. For OpenAI, that’s essentially a firehose of compute from the world’s most valuable chipmaker; for NVIDIA, it’s a way to lock in the most prominent AI customer on the planet, on top of its deep ties with Microsoft and other hyperscalers.

Then the wobble hit. In late January, The Wall Street Journal reported that NVIDIA’s plan to invest up to $100 billion in OpenAI had stalled amid internal concerns about the sheer scale and structure of the deal. People inside the company, according to those reports, worried about tying up so much capital in a single partner, especially as competition from Google, Anthropic, Meta, and a growing list of startups accelerates. The Journal also said Huang had privately played down the commitment, describing the 100 billion figure as non‑binding and suggesting the deal wasn’t locked.

That’s where the “unhappy with OpenAI” narrative came from: the idea that NVIDIA had gotten cold feet about betting so heavily on one AI lab whose strategy, governance, and appetite for cash have all been questioned over the last couple of years. Add in OpenAI’s complicated governance saga, its parallel partnerships with Microsoft and AMD, and a broader AI funding race where valuations are soaring, and it was easy to imagine tensions behind the scenes.

Publicly, Huang is doing damage control. Speaking to reporters in Taipei, he called it “nonsense” to say he was unhappy with OpenAI, and went out of his way to praise both the company and Sam Altman personally. “We are going to make a huge investment in OpenAI,” he said, describing it as “one of the most consequential companies of our time” and saying he “really love working with Sam.” He also stressed that OpenAI’s new funding round is being closed by Altman, and that NVIDIA will “absolutely be involved” and likely make the largest single investment in its history.

At the same time, Huang is nudging expectations down from the headline number. Asked specifically whether NVIDIA would still be investing more than $100 billion, he answered, “No, nothing like that,” while still using “huge” to describe what’s coming. That’s a neat bit of line‑walking: keep the relationship narrative positive, but quietly reframe the deal as big, not ridiculous.

Underneath the PR, there’s a real tension between hype and discipline. NVIDIA is already the default pick‑and‑shovel provider for generative AI, selling out its highest‑end GPUs as fast as TSMC and its partners can make them. Dedicating tens of gigawatts of capacity and a giant equity check to one customer isn’t just about friendship with OpenAI; it’s a way to secure long‑term demand, influence the direction of future models, and keep rivals from locking in the same supply. But from a shareholder perspective, it’s fair to ask whether an investment on that scale should look more like a carefully staged series of smaller commitments tied to concrete deployments, rather than a single headline‑grabbing figure.

OpenAI, for its part, has every incentive to keep NVIDIA close while also diversifying. It already has a massive cloud and equity relationship with Microsoft, plus an expanding partnership with AMD that will see the rival chipmaker supply high‑performance GPUs and several gigawatts of compute for OpenAI’s next‑gen infrastructure. Reports have also flagged Amazon as another potential investor, with talks around up to $50 billion, underscoring how many tech giants want a piece of the ChatGPT halo. In that context, Huang’s insistence that he’s not unhappy sounds less like denial and more like a reminder: NVIDIA still intends to be a central pillar of OpenAI’s compute stack, even if the exact check size is getting renegotiated.

There’s also a geopolitical and supply‑chain angle to Huang’s public charm offensive. His comments came during a high‑profile trip to Taiwan, where he met executives from TSMC and other major suppliers, urging them to boost output to keep up with AI demand. The “trillion‑dollar dinner” nickname wasn’t subtle: between NVIDIA, TSMC, and the rest of the table, you’re looking at a massive chunk of global market cap, all aligned around one big bet—that AI workloads will keep scaling and that customers like OpenAI will keep paying. If you’re Huang, you want to show those partners that big anchor customers are still locked in, even if the legal fine print is in flux.

Zoom out, and the whole episode is a good snapshot of where the AI industry is in early 2026: huge ambition, eye‑watering numbers, and a lot of improvisation. The NVIDIA–OpenAI relationship sits on top of a tangle of other mega‑deals—Microsoft’s multibillion‑dollar commitments, Oracle’s huge data‑center pact, and a fresh SoftBank‑led funding round that could value OpenAI in the hundreds of billions. Everyone wants compute, everyone wants capital, and everyone wants leverage, which is why a single phrase like “unhappy with OpenAI” can suddenly become a story big enough that the world’s most valuable chip CEO has to swat it down on a Taipei sidewalk.