What if the loudest voice against AI doom isn’t a philosopher with a Substack, but a guy in a leather jacket trying to move more chips?
I’ve spent the last two years reviewing AI tools for a living, which mostly means watching marketing decks collide with reality. So when people tell me Nvidia CEO Jensen Huang has become the great counterweight to AI doomerism, my first instinct isn’t to applaud. It’s to ask who benefits.
Let me be clear about what we actually know, because the gap between the verified record and the narrative is the whole story here.
What’s on the record
In January 2026, at CES in Las Vegas, Huang took the stage and introduced Alpamayo, an open-source AI platform for autonomous vehicles built to let cars reason through driving situations. Alongside it came a Mercedes-Benz partnership for a driverless car running on that platform, slated for a US launch first, then Europe and Asia.
He also talked money. A $35 billion circular investment arrangement with OpenAI. In March 2026, a projection of $1 trillion in AI chip sales, paired with talk about the company’s energy-efficient architecture. By September 2026, Nvidia’s board authorized another $150 billion in share repurchases. Huang picked up the IEEE Medal of Honor along the way. He’s also publicly criticized US tariffs and export rules, and pushed back on state-level AI regulation.
That’s the factual spine. Notice what’s missing from it: a single verified quote of Huang arguing that AI poses no existential risk to humanity.
The foil that assembles itself
The “Huang versus the doomers” framing isn’t something he had to declare. It’s something the market built around him by watching what he does.
Think about the actual content of those announcements. An open-source driving platform. A production car with a legacy automaker. A chip roadmap with a trillion-dollar ceiling. Buybacks at a scale that implies serious confidence in forward earnings. Every one of those moves is a bet that AI is a normal, enormous, boring industrial transition, the kind that needs supply chains and regulatory clarity and energy budgets.
Doomerism, by contrast, needs a discontinuity. It needs the moment where the thing we built stops being a product and starts being an agent with its own plans. You cannot build a ten-year fab and auto-partnership strategy around that premise. The two worldviews aren’t just rhetorically opposed, they’re operationally incompatible.
So Huang doesn’t have to argue against the apocalypse. He just has to keep scheduling product launches for 2027.
Why I’m not fully buying it
Here’s where my reviewer brain kicks in. The person selling the picks and shovels is the single worst-positioned individual to assess whether the mine collapses.
That $35 billion circular investment with OpenAI deserves a hard look. Circular means money flows out to a customer who flows it back as purchases. That’s a structure that can look like explosive demand right up until it looks like an accounting hall of mirrors. I’m not alleging anything improper. I’m saying anyone treating Nvidia’s revenue curve as independent evidence about AI’s trajectory should understand that part of that curve is Nvidia funding its own buyers.
Same with the regulatory pushback. Opposing state AI rules might reflect a genuine read that the rules are badly written. It also happens to be the position that keeps the largest possible market open. Both can be true simultaneously, and you should hold both in your head.
What this means for people actually using the tools
I review agents and AI products. Almost none of my work involves existential risk. It involves whether a product does what its landing page claims, and the answer is frequently no.
The doom debate and the hype debate share a flaw. Both argue about a hypothetical future system while ignoring the mediocre present one. An autonomous vehicle platform shipping to real roads with Mercedes-Benz is a testable claim. Either those cars handle edge cases or they don’t. That’s a far more useful conversation than whether a model might one day develop intentions.
If Alpamayo works, Huang’s implicit worldview gets strong supporting evidence, not because he won an argument but because the thing shipped. If it stalls, the trillion-dollar projection starts looking like a forecast built on forecasts.
My honest read
Huang is a useful corrective to a discourse that had drifted into pure speculation. He talks about energy efficiency, manufacturing, and launch windows while others talk about probabilities of extinction. That grounding has value.
But he’s not a neutral party, and treating a chip vendor’s optimism as philosophical refutation is lazy thinking. Watch the Mercedes cars. Watch whether that OpenAI money keeps circling. Judge the claims against shipped products, which is the only standard that’s ever told me anything true about this industry.
🕒 Published: