NVIDIA Without Jensen

Today, 12 November 2026, Jensen Huang resigned as NVIDIA’s CEO and board member. CFO Colette Kress is taking over as interim CEO while the board looks for a permanent successor. The news followed reports of two affairs, one with a former Megaspeed executive described as a suspected Chinese intelligence asset, and another with an NVIDIA employee. That intelligence allegation remains unproven publicly. NVIDIA’s board said it had no reason to believe national security or company secrets had been compromised, while opening an independent investigation. (NVIDIA’s transition announcement; NVIDIA board statement; reporting on the relationships)

It all happened so quickly that I felt a bit sick watching it unfold. It feels like a coup, like someone’s out for him. Maybe the board simply decided that, after the reports, he couldn’t stay. If so, that’s an insanely stupid conclusion, the kind of move people make when protecting their own power matters more than protecting the company. It’s why founder shares and voting control matter: a board shouldn’t be able to push out the person who built the company over a few days of headlines. I still catch myself hoping this is somehow temporary and he’ll be back once the story settles. I know that’s probably wishful thinking.

The affairs are his responsibility. I’m not going to defend them. But the thing I keep thinking about is what NVIDIA is without Jensen. He co-founded the company in 1993 and has led it through a series of bets that looked strange before they looked obvious. CUDA opened GPUs up to science and research. Then came machine learning, simulation, robotics and the broader AI computing platform. NVIDIA’s own history is basically a record of the company changing what its chips are for. (NVIDIA’s corporate timeline)

NVIDIA is also one of the few American companies putting serious weight behind open models and tools. Its open Nemotron models and robotics workflows give researchers and developers more to build on than a closed API. NVIDIA’s agreement to acquire Hugging Face is a direct bet on open source. Hugging Face is one of the main places developers share open models, and NVIDIA plans to put more resources behind it. Open source needs serious infrastructure and distribution. NVIDIA is in a position to help provide both. (NVIDIA on the Hugging Face deal; SEC filing on the deal; NVIDIA’s open-model work)

The same is true of compute. Jensen saw that the hyperscalers would build their own infrastructure, and backed neoclouds like CoreWeave and Nebius that can bring compute online outside Amazon, Microsoft and Google. NVIDIA has also helped neoclouds build their infrastructure businesses. Open source gives people options only if they have real choices about where to run it. Without neoclouds and other independent compute providers, the hyperscalers could still control access to the hardware underneath. Then the choice would be an illusion. (NVIDIA’s earnings-call transcript)

That’s why this scares me. NVIDIA is too important to the future of AI to become a company that only ships the next chip and protects its margins. It needs to support the entire ecosystem outside the mega AI labs and hyperscalers: open-model developers, independent researchers, robotics teams and neoclouds. Colette Kress is a finance executive, and right now she’s interim. I hope she stays interim. I pray Jensen comes back. To me, CFOs are the people least likely to make the kinds of bets Jensen did: backing CUDA, machine learning, robotics, open models and neoclouds before the numbers made the case. If Jensen doesn’t return, NVIDIA needs an engineer and a visionary at the helm, not an operations or finance person.

Jensen is NVIDIA. NVIDIA is Jensen. He co-founded it in 1993 and spent 33 years seeing around corners. CUDA, machine learning and accelerated computing were not separate bets to him; they were parts of the same future for the company. He kept pulling NVIDIA toward the next thing before most people could see it. No one else has his instinct or his record. You can hire another CEO. You can’t hire another Jensen.

I hope NVIDIA keeps taking those kinds of risks. I hope the research, open-source work and neoclouds stay central. And, selfishly, I hope this isn’t the last chapter of Jensen at NVIDIA. The company will survive without him. The real question is whether it can keep looking as far ahead.

I’ll leave you with a quote:

The best way to predict the future is to invent it.

Alan Kay (source)