Davos loves a story that sounds like inevitability. Jensen Huang showed up on the main stage with one that sounds like cold, hard concrete.

The Nvidia $NVDA CEO called AI “the largest infrastructure buildout in human history,” then built an argument sturdy enough for the room he was in: executives who approve budgets, policymakers who approve permits, and investors who approve the approving. He made a technology story sound like a construction story, and a construction story sound like a jobs story, and a jobs story sound like a pension-fund story.

In Huang’s 2026 version of the story, AI has already graduated from novelty to utility — the kind of utility that needs power, people, and a lot of approvals before anyone can actually get to the fun part.

The point is the buildout — and the bill

Huang framed AI as “a five-layer cake,” and the first layers are the ones that might make Davos people blink because they feel less like software and more like national plumbing: energy, “chips and computing infrastructure,” “cloud data centers,” “AI models,” and finally, the application layer. Every layer needs to be built, operated, maintained, secured, and expanded. Each layer brings its own bottlenecks, vendors, and political problems. Each layer also brings a bill.

The top — where hardware turns into margins — is “ultimately” the one where “where economic benefit will happen,” Huang said, describing a world where AI shows up inside financial services, healthcare, manufacturing, and whatever else still has inefficiencies to squeeze.

Larry Fink, the CEO of BlackRock $BLK, sitting beside Huang, summarized the mood the way a capital allocator does. Basically, if this is an AI bubble, it’s a strange one — the kind where the constraint looks less like demand and more like build capacity. Fink’s question landed on the practical anxiety that haunts every Davos year: Are we investing enough?

The “AI jobs” story he’s selling is blue-collar

Huang also spent time on the part of the AI economy that rarely gets a spot in a demo. He highlighted demand for “plumbers, electricians, construction workers, steelworkers, network technicians” and the teams that install and operate advanced equipment. You can hear the strategic advantage in making a technology boom sound like a labor boom that doesn’t require everyone to learn Python.

The rooms in Davos contain plenty of people who worry about social backlash the way they worry about inflation, as a force that arrives suddenly and ruins the mood. A jobs story anchored in trades gives AI a sturdier public-facing shape. Huang is suggesting that a platform shift that needs physical buildout spreads the payroll beyond engineers and venture-backed founders. Every layer of the stack needs people who can build and run it. Energy systems get expanded. Facilities get wired. Racks get installed. Networks get maintained. The “cloud” keeps behaving like a warehouse with a utility bill.

Huang also pulled the conversation upward to the startup layer, saying 2025 was among the biggest years for VC funding on record, with most capital flowing into “AI-native companies,” and he treated that cash as downstream job creation. Capital builds the application layer. Infrastructure and labor follow.

AI won’t erase work; it’ll reassign it

Huang joked that watching him and Fink, “You would probably think the two of us are typists.” Fine — automate the typing. The job still exists. The deeper point is the one Huang wanted to travel beyond Switzerland — and one he’s made before: “So the question is, what is the purpose of your job?”

He used healthcare as Exhibit A. AI, he said, is “a key tool in radiology,” and yet there are “more radiologists than ever.” He added, “If you reason from first principles, not surprisingly, the number of radiologists has gone up,” claiming that AI accelerates the scan-reading task and frees people to do the higher-stakes parts.

The U.S., Huang said, faces “a shortage of roughly five million nurses,” partly because “nurses spend nearly half their time on charting and documentation.” He pointed to the use of AI for charting and transcription and claimed the productivity loop ends with more hiring. “Surprisingly — or not surprisingly,” he said, “AI is increasing productivity, and as a result, hospitals want to hire more people.” The room heard what it wanted to hear: the idea that AI expands capacity, improves outcomes, and keeps the labor story calm.

“AI is infrastructure” is also a sovereignty pitch

Huang kept returning to the word “infrastructure.” “AI is infrastructure,” he said. “You should have AI as part of your infrastructure.” Then, he urged countries to build their own capabilities and keep them close. “Develop your AI, continue to refine it, and have your national intelligence be part of your ecosystem,” he said, emphasizing local “language and culture.”

He also leaned hard on accessibility, a crucial ingredient if you want “infrastructure” to sound like a public good. “AI is super easy to use — it’s the easiest software to use in history,” he said, noting that in “two to three years,” AI tools have reached “nearly a billion people.” He treated that speed as inevitability driven by usability, then defined the new baseline skill set. “It is very clear that it is essential to learn how to use AI — how to direct it, manage it, guardrail it, evaluate it.”

Europe’s assignment: turn industrial strength into robots

Europe’s strength, Huang suggested, sits in industrial systems and manufacturing depth — precisely the terrain where “physical AI” and robotics become more than a talking point. “You don’t write AI — you teach AI,” he said, adding that “robotics is a once-in-a-generation opportunity,” especially for nations with strong industrial bases. If Europe wants a starring role in the AI buildout, Huang’s view places it in the physical world — factories, machines, industrial systems that learn.

Fink closed the loop the way Davos likes best, with ownership, which sits underneath every Davos conversation. “I actually believe it’s going to be a great investment for pension funds around the world to be a part of that, to grow with this AI world,” he said. “We need to make sure that the average pensioner and the average saver is part of that growth.” Huang didn’t resist the ending. “The opportunity,” he said, “is really quite extraordinary, and everybody ought to get involved.”

In Davos, that’s the cleanest kind of promise: The buildout is massive, the work is widespread, the politics are manageable, and the upside has room for everyone. The room, predictably, loved it.

📬 Sign up for the Daily Brief

(责任编辑：)