Speaking at Carnegie Mellon's commencement in May 2026, Nvidia CEO Jensen Huang told graduates that AI is creating opportunity far beyond software and engineering degrees, naming electricians, plumbers, ironworkers, technicians, and builders directly. "AI gives America the opportunity to build again," he said. "Electricians, plumbers, iron workers, technicians, builders, this is your time." The claim isn't just a soundbite: independent labor data shows trades demand growing roughly three times faster than office roles since 2022. It's also not disinterested. Nvidia's entire business depends on the physical infrastructure those trades workers build.
What Huang actually said
Huang delivered the keynote at Carnegie Mellon's 128th commencement on a rainy Sunday morning in Pittsburgh, addressing roughly 5,800 graduates. He received an honorary doctorate and used the speech to frame the current moment as a full reindustrialization of the American economy, not just a software revolution. "AI is not just creating a new computing industry," he told the crowd, "it is creating a new industrial era."
The trades reference wasn't a throwaway line. Huang explicitly listed electricians, plumbers, ironworkers, technicians, and builders as direct beneficiaries of the AI buildout, standing alongside the computer science and robotics graduates he was there to honor. Carnegie Mellon was a deliberate choice of venue: the university's School of Computer Science created the Logic Theorist in the 1950s, widely regarded as the first AI program, and founded the world's first academic robotics institute in 1979.
Why a chip company CEO is talking about plumbers
The connection is direct, not sentimental. Nvidia sells the processors that power AI systems, but those processors are worthless without somewhere to run. Every data center needs electrical capacity measured in megawatts, industrial cooling systems running continuously, and physical construction before a single chip generates revenue, a dynamic covered in more detail in our reporting on data center facility technicians and the broader skilled trades shortage.
Capital expenditure from the largest technology companies could reach roughly $700 billion in 2026 alone, largely driven by data center construction. Every dollar of that spending eventually needs electricians, HVAC technicians, and ironworkers to actually build the thing being paid for. Huang's business grows only as fast as that physical infrastructure gets built, which makes trades labor a genuine bottleneck for Nvidia's own growth, not just a talking point.
The data behind the speech
Independent of Huang's framing, the underlying numbers hold up. A Randstad analysis of more than 150 million job postings found skilled trades demand growing at roughly three times the rate of desk-based roles since 2022. Construction worker postings climbed 30 percent, welder postings rose 25 percent, and electrician postings increased 18 percent over the same period. Demand for robotics technicians, a newer and smaller category, surged 107 percent.
This lines up with the broader shortage numbers covered in our earlier reporting on the corporate push into trades training: multiple large companies, not just Nvidia, have independently concluded that trades labor is the binding constraint on their own growth plans.
Worth a grain of skepticism
None of this means Huang's framing should be taken uncritically. He has an obvious and direct financial interest in the AI infrastructure buildout being seen as broadly beneficial rather than narrowly enriching a small group of technology companies. Telling a graduating class that trades work is dignified and valuable is also, conveniently, a message that keeps public sentiment favorable toward the massive capital spending Nvidia depends on.
Huang has also used a specific rhetorical move repeatedly in recent speeches: distinguishing between a job's "task" and its "purpose," arguing AI automates the task while elevating the purpose. It's a clean line, and it holds up in some settings better than others. The honest version of this story is that the trades data is real and independently verifiable, while the framing around it serves Nvidia's interests as much as it serves any electrician's.
What this means practically
Set aside who's saying it, and the underlying case for trades work in 2026 doesn't actually depend on Jensen Huang's endorsement. It rests on the same numbers covered throughout our reporting: real BLS wage data, real job posting growth, and a real, well-documented labor shortage. Huang's speech is a useful marker of how mainstream that view has become, a Fortune 500 tech CEO now says this from a commencement stage, but it's a symptom of the trend, not the source of it.
- What did Jensen Huang say about electricians and plumbers?
- Speaking at Carnegie Mellon University's commencement in May 2026, Nvidia CEO Jensen Huang told graduates that AI is creating demand far beyond software and engineering, naming electricians, plumbers, ironworkers, technicians, and builders directly as entering a new industrial era alongside computer scientists.
- Why would the CEO of an AI chip company care about electricians?
- AI data centers require enormous physical infrastructure to build and run: electrical capacity, industrial cooling, and construction labor. Nvidia's chips only generate revenue once that physical infrastructure exists, so demand for trades workers is directly tied to how fast Nvidia's own market can grow.
- Is the trades demand from AI data centers backed by real data?
- Yes. A Randstad analysis of more than 150 million job postings found skilled trades demand growing roughly three times faster than office roles since 2022, with construction postings up 30%, welder postings up 25%, and electrician postings up 18%.
- Should Jensen Huang's comments be taken at face value?
- The underlying trades demand is real and independently documented. But Huang has a direct financial interest in the AI infrastructure buildout succeeding and being seen as broadly beneficial, which is worth factoring in when weighing how much credit to give the framing itself versus the data behind it.