“Without These Workers, America Can’t Compete with China in AI”, Meta‘s President Just Named the Workforce Crisis No One Saw Coming

America was supposed to be winning the AI race.

For years, the narrative was clean: the United States had the best labs, the biggest budgets, and the brightest minds. OpenAI, Google DeepMind, Anthropic, all American. And China? They were chasing.

Then something shifted.

Not in a flashy, headline-grabbing way. No single model toppled GPT overnight. But slowly, methodically, the gap began to close. By early 2025, Chinese models were trading places with American ones at the top of global leaderboards. DeepSeek-R1 marked the first large instance where a Chinese model pulled ahead. As of March 2026, the performance gap between US and Chinese AI models had shrunk to just 2.7%.

Here‘s what nobody told you about that gap: It’s not closing because Chinese chips suddenly got faster.

It‘s closing because China built a workforce machine. And America can’t keep up.

That brings us to Dina Powell McCormick, Meta‘s President and Vice Chairman. Last week, she sat down with Maria Bartiromo on Fox Business and said something that should worry every American who cares about the future of technology. Something that has nothing to do with billion-parameter models or GPU clusters. Something far more human.

She said: ”If the country, if America doesn‘t come together and ensure that we frankly treat these workers as the American heroes that they are, without them, we can’t compete with China.“*

Electricians. Fiber technicians. Welders.

That‘s who she’s talking about.

So let me ask you something. When you think about the AI arms race, the battle to build machines smarter than any human, do you picture a guy in a hard hat running conduit through a half-constructed building?

Probably not.

But maybe you should. And after you finish this article, I think you‘ll see why.

Let’s unpack what McCormick actually said, why these workers matter more than you think, and what‘s at stake if America keeps ignoring the most basic ingredient of the AI revolution: skilled hands to build the physical infrastructure that powers it all.

---

What Meta’s President Actually Said (And Why It Matters)

Here‘s the context you won’t get from a two-sentence news blurb.

Meta is spending billions on AI infrastructure. New data centers, expanded server farms, fiber networks crisscrossing the country. But there‘s a bottleneck that has nothing to do with NVIDIA H100 chips or large language model hallucinations. McCormick put it plainly: Tech companies racing to expand artificial intelligence infrastructure are increasingly running into a challenge that has little to do with software or chips, finding enough skilled workers to build and maintain the facilities powering the AI boom.

She didn’t pull this observation out of thin air. Alongside Mike Rowe, yes, that Mike Rowe, the Dirty Jobs guy whose foundation has spent years promoting skilled trades, McCormick announced America‘s Workforce Academy, a new training initiative launching in Louisiana, Ohio, Indiana, and Texas. The program offers paid training, industry-recognized credentials, and upon completion, a guaranteed job opportunity.

A couple of details worth pausing on:

• Paid training. You don’t have to quit your current job to participate. That matters, a lot, because the people who might fill these roles can‘t afford to take six months off to learn a new trade.
• Guaranteed job opportunity. Not “maybe we’ll call you.” A real commitment that someone will be waiting on the other side.

Rowe added something telling on the same broadcast: ”The jobs we‘re talking about, by and large, exist out of sight and out of mind.“*

Out of sight. Out of mind. And yet, absolutely essential.

If you take nothing else away from this section, take this: The AI revolution doesn’t just happen in clean rooms staffed by hoodie-wearing coders. It happens on construction sites. It happens at welding stations. It happens in the crawl spaces beneath data centers where fiber optic cables are spliced together by hand.

Those workers are not a sideshow. They‘re the show.

---

The Workers Nobody Is Talking About

Let’s get specific about who McCormick means.

Electricians

Data centers consume staggering amounts of electricity. We‘re talking about facilities that require the same power draw as a small city. You can’t just plug them into the nearest outlet. The electrical systems need to be designed, installed, maintained, and, inevitably, repaired. Someone has to do that work. Someone skilled has to do that work. And right now, we don‘t have enough someones.

Fiber Technicians

AI doesn’t work without connectivity. Models need to communicate. Data needs to move. Fiber optic networks are the nervous system of the AI economy. Fiber technicians are the people who install, splice, test, and maintain those networks. It‘s delicate work. A single bad splice can take down connectivity for an entire facility. And again: we don’t have enough trained people to do it.

Welders

You don‘t think about welding when you think about AI. But data centers are made of steel, massive steel frames, cooling systems, structural supports. Welders fabricate and assemble those components. The work requires precision, experience, and a certification that takes years to earn. McCormick named welders explicitly, alongside electricians and fiber technicians. Those three roles form the backbone of physical AI infrastructure.

The United States had a shortage of electricians, welders, and other skilled manual laborers before Meta and its rivals embarked on their AI buildouts. Now demand is accelerating. Supply is not.

Let that sink in. We‘re trying to build the most technologically advanced infrastructure in human history, and we don’t have enough people who know how to bend conduit.

---

The Numbers Don‘t Lie: America Is Losing the Worker War

I’m going to give you a set of statistics. Try not to glaze over. Each one is a warning siren.

10-to-1. China now graduates roughly 1.3 million engineers per year. The United States graduates roughly 130,000. That‘s not a gap. That’s a chasm.

What does that gap actually mean? In practical terms, it means longer development cycles. It means product improvements get deferred. It means requisitions for engineering talent go unfilled. When you‘re racing against a country with ten times your output of technically trained graduates, you’re not just losing slowly. You‘re losing exponentially.

80% drop. According to the Stanford AI Index 2026, despite massive US investments, the country is struggling to attract global talent, with an 80% drop in AI researchers and developers choosing to move to the United States. The brain drain has reversed direction. And here‘s the kicker: many of those researchers are Chinese-born professionals who would have stayed in America a decade ago.

38% to 12%. The Economist tracked researchers publishing at NeurIPS, the world’s premier AI conference. In 2019, 29% of researchers at the conference had early-career origins in China. By 2025, that figure had climbed to 50%. Meanwhile, the share with early-career origins in the US dropped from 20% to 12%.

2.7% That‘s the remaining performance gap between US and Chinese top-tier models. As of March 2026, it’s effectively closed.

The Stanford AI Index doesn‘t sugarcoat it: ”The US is no longer leading the AI model race.“*

When the people who build the models stop coming to your country, and the people who build the infrastructure don’t exist in sufficient numbers, and the performance gap between your models and your competitor‘s models is statistically negligible, what exactly are you still leading?

---

How China Built a Talent Machine, And Why It’s Winning

China‘s approach to AI talent isn’t accidental. It‘s systemic.

The Domestic Pipeline

China has become the second-largest AI talent pool in the world, with AI researcher numbers growing from roughly 4,400 in 2015 to about 52,000 in 2024, a 28.7% compound annual growth rate. By 2025, China had 785 vocational colleges with AI technology programs. They are producing talent at scale, from elite research PhDs to hands-on technical workers.

The Reverse Brain Drain

For years, Chinese AI talent left and never came back. In 2019, only 12% of Chinese AI researchers who earned graduate degrees overseas returned to China. By 2025, that figure had risen to 28%. Multiple factors are driving this shift: tightening US visa policies, the rise of competitive Chinese LLMs, and significant investment in domestic research infrastructure.

Put simply: America used to import China‘s best brains and keep them. Now, more of those brains are going home.

The Results Are Visible

DeepSeek, the Chinese AI company that briefly knocked OpenAI off the top of the app store charts, is “essentially a domestic talent success story.” Half of DeepSeek’s team never studied abroad. Those who did came back. Stanford‘s report noted that China doesn’t view overseas study as talent loss, it views it as a knowledge acquisition strategy, with the expectation that people will eventually return and apply what they learned.

You don‘t beat that system with signing bonuses alone.

---

The Immigration Paradox: Blocking the Brains We Need

Here’s where the story gets uncomfortable for anyone who likes simple answers.

The United States has an immigration problem. Not too many immigrants, too few of the right immigrants, arriving via systems that are broken.

Consider this: 77% of the top AI companies in the United States were founded or co-founded by first- or second-generation immigrants. Elon Musk. Jensen Huang. The founders of Google, Microsoft, OpenAI, immigrants or children of immigrants. The American AI boom is an immigrant story.

And yet, tightening US immigration policies appear to be stinging Silicon Valley as top American tech giants face a trend of brilliant Chinese scientists increasingly opting to return home, taking Washington‘s ambitious AI visions back to Beijing.

According to the National Foundation for American Policy, 85 to 90% of Chinese students studying computer science would stay in the US if they could. The problem isn’t willingness. The problem is the system. H-1B caps, visa uncertainty, and a political climate that often treats skilled immigration as a threat rather than an asset.

The AI race has shifted from a battle over microchips to a war for top AI talent. Which side is making it easier for that talent to stay?

China, that‘s which side.

While the US has tightened restrictions, China has launched visa programs designed to attract young professionals with STEM degrees. They’re not just competing with American salaries anymore. They‘re competing with American access, and they’re winning.

---

Meta‘s $14 Billion Bet (And Why It Needs Your Local Electrician)

If you want proof that McCormick isn’t just talking, look at Meta‘s checkbook.

In June 2025, Meta invested roughly $14 billion to acquire Scale AI and brought its founder, Alexandr Wang, on board. That’s not a small bet. That‘s the kind of money that rearranges industries.

Meta’s new superintelligence lab, the one tasked with building AI that surpasses human intelligence, has a team composition that tells its own story. Of 11 publicly listed hires, 7 are ethnically Chinese (and that‘s not counting the CEO and former GitHub CEO brought in to lead the unit). Four of the new hires are reportedly from Tsinghua University, China’s MIT.

McCormick said salaries for top AI talent have reached $100 million signing bonuses, and she expects them to go higher, $200 million range within a few years.

You can offer $200 million to a handful of elite researchers. But you cannot manufacture welders with signing bonuses.

You cannot conjure fiber technicians with stock options.

And you cannot win the AI race if your data centers are sitting half-built for lack of skilled labor.

That‘s the paradox at the heart of McCormick’s message. America is building the world‘s most advanced computational infrastructure on a foundation of workers who are, in Rowe’s words, out of sight and out of mind.

---

What’s at Stake If America Doesn‘t Act

Let me be direct with you.

Economically: The US currently produces more top-tier AI models and attracts the lion‘s share of private investment. But China leads in industrial robot installations, patent output, and, crucially, the volume of AI talent it produces. If the US loses its ability to field enough engineers and enough skilled trades workers simultaneously, its economic advantage evaporates.

Geopolitically: This isn’t about tech bragging rights. The nation that leads in AI sets the standards for the global economy, trade rules, security frameworks, ethical norms. Losing that position doesn‘t mean second place. It means playing by someone else’s rules.

For workers: The most immediate consequence is already visible. The US Bureau of Labor Statistics projects a shortage of 1.2 million tech workers by 2026. AI-related jobs are growing at twice the rate of the overall workforce. Meanwhile, entry-level positions for young developers (ages 22-25) have fallen nearly 20%.

The message for a young person entering the workforce is confusing at best. ”We need more tech workers. Also, entry-level tech jobs are disappearing. Also, there are high-paying skilled trade jobs nobody told you about.“*

That confusion is a policy failure. It‘s also an opportunity, if America can get its act together.

---

How America Wins the Worker War

This isn’t a hopeless situation. But it does require action. Here‘s what winning looks like.

For Policymakers:

• Fix the immigration system. Raise H-1B caps. Create green card exemptions for advanced STEM degree holders. Protect OPT and STEM OPT programs. The talent is willing to come. Stop blocking the door.
• Invest in vocational training at scale. America’s Workforce Academy is a start. But one program in four states isn‘t going to close a million-person gap.
• Reform the narrative around skilled trades. For two decades, American students have been told that the only path to success is a four-year degree. That message has created a shortage of people who can actually build things. It’s time for a correction.

For Employers (including Meta):

• Expand apprenticeship models. Paid training with guaranteed placement is the right formula. Do more of it.
• Partner with community colleges and trade schools. Don‘t wait for the education system to fix itself. Build the pipeline you need.
• Offer career pathways, not just jobs. Retention matters as much as recruitment. People stay where they see a future.

For Individuals: If you’re a young person (or someone considering a career change) and you‘re reading this, here’s something to consider: Skilled trades pay well. Electricians, welders, and fiber technicians earn solid middle-class wages. The work is hands-on, tangible, and increasingly essential to the most exciting technological revolution in a century. McCormick called these workers American heroes. That‘s not just rhetoric. It’s recognition of what the market is already signaling.

If you‘re already in a skilled trade, the AI boom represents a massive tailwind for your career. The people who know how to build and maintain physical infrastructure are about to become very, very valuable.

If you’re a policymaker or business leader reading this, the window for action is closing. China isn‘t waiting. Every year the US fails to address its workforce gap, China’s advantage compounds.

---

The Ending 

Meta‘s president didn‘t say America can’t compete with China in AI because of chips. She didn‘t say it’s because of algorithms or investment or research output.

She said it‘s because of workers.

The workers who build the data centers. The workers who run the fiber. The workers whose names you’ll never see on a byline or a research paper, but whose absence will be felt the moment a project stalls for lack of skilled hands.

America has spent the last decade marveling at its own technological brilliance. Silicon Valley has been drunk on its own mythology. And while we were busy congratulating ourselves, China quietly built a workforce machine that now threatens to overtake us in the one area where we thought we had an unassailable lead.

The good news? We can still fix this.

But fixing it requires recognizing a simple truth: The AI revolution isn‘t just a story about brains. It’s a story about hands. And until America starts treating the people with those hands like the heroes they are, we‘re going to keep losing ground to a country that already understands this.

”If America doesn‘t treat these workers as the American heroes that they are, without them, we can’t compete with China.“ , Dina Powell McCormick, President of Meta

 The race for AI dominance has never been just about who has the smartest algorithms. It‘s about who can build the physical infrastructure to run those algorithms at scale. It’s about who can attract and retain the human talent, both the PhD researchers and the certified welders, to make the whole system work.

Meta‘s president just handed America a warning. The question is whether we’re willing to hear it.

Right now, somewhere in Louisiana, Ohio, Indiana, or Texas, a worker is going through paid training at America‘s Workforce Academy. That person is going to walk out with a credential, a guaranteed job opportunity, and a role in building the AI infrastructure that will define the next decade.

That’s a start.

But it‘s not enough.

We need more programs like that. We need immigration policies that welcome the world’s best minds instead of turning them away. We need a national conversation that treats skilled trades as careers of choice rather than fallback options. And we need to recognize, finally, that the people who bend conduit and splice fiber are not optional extras in the AI revolution. They are the revolution.