By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panelists agree that the AI infrastructure buildout is facing structural bottlenecks, particularly in skilled trades and power/cooling. They disagree on the duration and sustainability of this boom, with some seeing it as a 3-5 year opportunity and others a decade-long cycle. The key risk is the normalization of AI capex post-2027, which could lead to a slowdown in trades demand and wage growth.
Risk: Normalization of AI capex post-2027
Opportunity: Utility-scale electrical infrastructure providers
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
When Jensen Huang showed up at Carnegie Mellon, the Nvidia CEO used his commencement address to the class of 2026 to make a case for a very different kind of career: one that requires a hard hat, not a laptop.
"AI gives America the opportunity to build again," Huang told the assembled crowd in Pittsburgh. "Electricians, plumbers, iron workers, technicians, builders — this is your time. AI is not just creating a new computing industry; it is creating a new industrial era (1)."
It's a striking message to deliver at one of the country's top computer science and engineering universities (2). But the data suggests Huang is onto something, and for anyone weighing their career options, the numbers are hard to ignore.
The scale of infrastructure being built to power artificial intelligence is staggering. Fortune reports that capital spending from the largest U.S. tech companies could hit $700 billion this year alone, driven by data center construction and the infrastructure needed to train and deploy AI models (3).
Worldwide, McKinsey projected the data center boom could generate close to $7 trillion in investment by 2030 (4).
None of that gets built without people swinging hammers, pulling wire and laying pipe.
Randstad's March analysis of more than 150 million U.S. job postings found that demand for skilled trades is now growing three times faster than for professional desk-based roles (5).
Since generative AI hit the mainstream in late 2022 (6), postings for construction workers, welders and electricians are up between 18% and 30%. More specialized roles have seen even sharper spikes: demand for robotics technicians jumped 107%, and HVAC engineers surged 67%.
"AI can't build data centers, upgrade power grids, or maintain its own infrastructure," Greg Dyer, Randstad North America's chief commercial officer, said in the report (7).
Meanwhile, the pipeline of workers entering the trades isn't keeping up. Companies can't hire fast enough to replace the waves of older tradespeople now entering retirement — Randstad found for every 100 young workers entering manufacturing, 102 are leaving (8).
Huang put it simply in his speech: "This is the largest technology infrastructure buildout in human history and a once-in-a-generation opportunity to reindustrialize America (9)."
Read More: From $10 to $100K: Build your real estate empire using the same strategies as the 1%
Here's what makes Huang's message relevant beyond the rhetoric: the pay in the skilled trades has become competitive with many college-degree roles, but without the debt.
According to the U.S. Bureau of Labor Statistics (BLS), the median annual wage for electricians was $62,350 as of May 2024, with employment projected to grow 9% through 2034 — much faster than the average across all occupations (10).
The BLS also reports median pay for plumbers, pipefitters and steamfitters was $62,970 (11). The top 10% of electricians earn more than $106,030 (12) annually; for plumbers, that figure exceeds $105,150 (13).
Those earnings come through apprenticeship programs, not four-year degrees. Electricians, for instance, typically complete a four- to five-year apprenticeship (14) that includes paid on-the-job training from day one. As Fortune noted (15),(16), these paths let workers earn while they learn — offering a faster, often cheaper, route into the workforce than a traditional college education.
Huang has previously suggested that trades workers could soon command six-figure salaries even early in their careers, according to Fortune (17). And CNBC reporting backs that up: advertised wages for HVAC engineers have risen 10–15% over the past four years, and specialists moving into high-level data center roles can see pay bumps of 25–30% (18).
The opportunity is real, but it isn't guaranteed, as surge in skilled trade demand is heavily concentrated in data center construction, and not all signals in the broader industry are positive.
Fortune notes that data center construction actually slowed last year for the first time since 2020, as developers hit delays in zoning, permitting and securing adequate power supply (19).
Private nonresidential construction contracted in six of the seven months through mid-2025, per ABC data (20), and tariff-related disruptions and immigration enforcement actions have compounded the pressure: 28% of contractors reported related workforce impacts, ENR reports (21).
"With the exception of the ongoing boom in data center construction…there are few sources of momentum," Anirban Basu, Chief Economist at the Associated Builders and Contractors, noted (22).
And trades workers building data centers don't necessarily get permanent employment once their portion of the work wraps up.
For a generation entering a labor market rattled by uncertainty, the signals are mixed but real.
A November 2025 Stanford study found a 16% decline in early-career employment across highly AI-exposed occupations since ChatGPT's launch, with employment among developers aged 22–25 dropping almost 20% from its late-2022 peak, Fortune reports (23).
The skilled trades, by contrast, offer something increasingly rare: work that AI genuinely cannot replicate, and that the AI era is actively making more valuable. As Randstad CEO Sander van't Noordende put it to Fortune: "AI is now revealing just how critical these roles are and how elevated they are becoming (24)."
Join 250,000+ readers and get Moneywise’s best stories and exclusive interviews first — clear insights curated and delivered weekly. Subscribe now.
We rely only on vetted sources and credible third-party reporting. For details, see our ethics and guidelines.
Fortune (1),(3),(9),(15),(16),(17),(19),(23),(24); CollegeVine (2); McKinsey (4); PR Newswire (5),(7); Randstad (6),(8); U.S. Bureau of Labor Statistics (10),(11),(12),(13),(14); CNBC (18); Construction Dive (20); ENR (21); Associated Builders and Contractors (22)
This article originally appeared on Moneywise.com under the title: Nvidia CEO tells college grads that electricians, plumbers and iron workers have a leg up – and there’s no degree needed
This article provides information only and should not be construed as advice. It is provided without warranty of any kind.
Four leading AI models discuss this article
"The AI revolution is currently constrained by physical labor and power infrastructure, creating a temporary, high-wage cycle for trades that is vulnerable to interest rate sensitivity and project-based volatility."
Jensen Huang’s pivot toward the trades is a masterclass in supply-side reality. While software eats the world, the $700 billion AI infrastructure buildout is hitting a hard physical ceiling: power and cooling. We are seeing a structural labor shortage in skilled trades that acts as a bottleneck for companies like NVDA and hyperscalers. However, investors should be wary of the 'duration' of this boom. Data center construction is lumpy and project-based; it lacks the recurring revenue profile of software. While the trades offer high utility and wage growth, the broader construction sector remains highly sensitive to interest rates and regulatory gridlock, which could throttle the very demand Huang is banking on.
The massive capital expenditure in data centers may lead to an 'overbuild' scenario, where the supply of compute capacity outstrips actual AI monetization, leading to a sudden, painful contraction in construction demand.
"N/A"
[Unavailable]
"The skilled trades surge is real but narrowly concentrated in data center construction facing permitting/power bottlenecks; once that wave crests, demand reverts to historical norms, leaving workers exposed to cyclical unemployment."
Huang's pitch is tactically smart but masks a structural fragility. Yes, skilled trades demand is up 3x faster than desk roles, and electrician/plumber wages ($62k median, $105k+ top decile) compete with many bachelor's-degree careers without debt. The $700B capex spend this year is real. BUT: the article itself admits data center construction slowed last year due to zoning/permitting/power constraints, and private nonresidential construction contracted in 6 of 7 months through mid-2025. These aren't cyclical dips—they're structural bottlenecks. Once the current buildout wave completes, demand normalizes. Trades workers face project-to-project volatility, not stable careers. Huang is selling a generational opportunity that may last 3–5 years, not a permanent shift.
If power grid modernization, reshoring, and data center density keep accelerating through 2030 (McKinsey's $7T thesis), trades workers could see sustained wage growth and genuine full-time stability—making this a real career inflection, not a temporary bubble.
"The core claim is that the no-degree trades boom is a durable macro shift, but in reality it hinges on a cyclical capex wave that could fade with higher financing costs, permitting bottlenecks, or a slower AI adoption."
Huang's blunt framing spotlights a cyclical, infrastructure-led AI buildout that could lift trades and nondegree paths. But the article glosses over risks: this appears to be a capex-driven surge rather than a structural labor reallocation. Data-center construction has shown signs of slowing due to zoning, power constraints, and permitting delays; private nonresidential work contracted in six of seven months through mid-2025; immigration and tariff frictions add friction to labor supply. A recession, higher financing costs, or slower AI adoption could invert the thesis. Investors should focus on beneficiaries of capex cycles—data-center builders and essential electrical/HVAC equipment—rather than assuming a durable, long-run shift toward no degree required.
But the capex boom could be front-loaded and not durable; once facilities are built, demand could normalize. In parallel, automation and outsourcing could dampen ongoing labor intensity long after installation.
"The grid modernization requirement creates a decade-long capital expenditure cycle that persists regardless of AI monetization success."
Claude and ChatGPT are missing the second-order effect of the 'power wall.' It isn't just about data center construction; it’s about the massive, multi-year grid modernization required to support these loads. This isn't a 3-5 year bubble; it’s a decade-long utility infrastructure upgrade cycle. Even if AI monetization stalls, the grid remains woefully inadequate for current electrification trends. Investors should look at utility-scale electrical infrastructure providers, not just the data center contractors themselves.
[Unavailable]
"Grid modernization is decoupled from AI capex cycles and won't sustain the wage premium Huang is selling once data center construction normalizes."
Gemini's grid modernization angle is real, but conflates two separate timelines. Utility capex (10+ years) and data center construction (3–5 years) operate on different cycles with different funding sources. Grid upgrades happen whether AI booms or busts; they're driven by EV adoption, electrification mandates, and aging infrastructure. The risk: Huang's trades pitch assumes *sustained* data center demand, not just grid work. If AI capex normalizes post-2027, trades workers face a decade of grid work at lower intensity—not the wage trajectory he's implying.
"Grid modernization is not a guaranteed, durable driver for trades; policy, permitting, and supply constraints could keep grid work volatile, limiting long-run wage stability even if AI capex remains buoyant."
Claude raises a valid long view, but the grid 'wall' argument risks becoming a tautology that just shifts the risk from data centers to utilities. In reality, grid modernization is subject to political cycles, permitting, and material/transformer shortages that can throttle pace for years, not decades. If AI capex cools post-2027, trades continue to be pulled by grid projects only if funding remains robust; otherwise, wage growth and full-time stability may prove episodic rather than structural.
The panelists agree that the AI infrastructure buildout is facing structural bottlenecks, particularly in skilled trades and power/cooling. They disagree on the duration and sustainability of this boom, with some seeing it as a 3-5 year opportunity and others a decade-long cycle. The key risk is the normalization of AI capex post-2027, which could lead to a slowdown in trades demand and wage growth.
Utility-scale electrical infrastructure providers
Normalization of AI capex post-2027