America Is Running the Wrong AI Race

— First to Scale Wins, Not First to Invent

by Alvin W. Graylin July 2025
Digital Fellow, Stanford Digital Economy Lab |
Author, Our Next Reality | Chairman, Virtual World Society

Imagine this: Two nations stand at the edge of a great technological chasm. One rushes to leap first, chasing glory and headlines. The other calmly builds a bridge, brick by brick, for its people to walk across safely.

Right now, America is the one sprinting toward the edge— where our government and our leading AI labs are obsessed with being the first to Artificial General Intelligence (AGI - AI smart enough to replace human cognitive labor). The amazing innovation engine in the U.S. has already brought the nation to the leading position in this key area. But are we running the wrong race?

According to the White House’s newly released “Winning the Race, America’s AI Action Plan”, the country is focused on accelerating innovation in order to “achieve and maintain unquestioned and unchallenged global technological dominance” and “achieve global dominance in AI”. In addition to an unfettered pursuit of raw speed of progress, the plan does contain important updated policy recommendations regarding the need for better energy infrastructure, advocating for open-source solutions, reducing vulnerabilities to bad actor misuse, and touches upon some limited programs for workforce protection. However, this new plan (as are prior AI policy papers) is still missing key elements of a holistic strategy needed to deliver long-term U.S. success.

Winning in AI isn’t about being first to build the smartest machine and dominance in a specific technology. It’s about being first to scale it wisely and broadly—across every school, hospital, factory, and business. It’s about ensuring our nation as a whole, not just a few tech companies, thrive in the AI era.

If we charge ahead without preparing the ground beneath us—without modernizing our institutions, protecting our population, upgrading our infrastructure, considering environment and AI safety concerns, and rebuilding trust domestically and abroad—we may win the moment but lose the century.

The Illusion of the First-Mover Advantage

History is littered with tech pioneers who beat others to market first…and then got buried. Friendster beat Facebook. AltaVista beat Google. Nokia beat Apple. But those early movers stumbled, while the fast-followers built empires.

Why? Because it’s not who invents first, it’s who deploys best. A study by organizational psychologist Adam Grant found that first movers fail nearly six times more often than those who wait, learn, and scale.

AGI will be no different, but the stakes are even higher. If we unleash world-changing technology before we have the social systems to absorb it, we’ll simply become the proving ground for others. China, Europe, or even rising nations in the Global South may leapfrog by broadly applying benefits of our breakthroughs while we wrestle with social backlash and instability at home.

America’s Social Infrastructure Is Not Ready

Our national focus has been locked on compute power, parameter counts, and GPU hoarding. But what about human dignity and social stability?

Our education system is still teaching kids to memorize facts in a world of instant answers. Our healthcare system is overwhelmed and overpriced. Our public transportation and broadband infrastructure still leave millions disconnected. And tens of millions of workers—from truck drivers to paralegals—face job displacement from automation without a plan.

The newly announced “AI Workforce Research Hub” in the Plan and expanded retraining efforts are a step in the right direction—acknowledging that job disruption is real. But we must ask: is it fast enough, and does it reach the workers who need it most?

This isn’t speculation. The IMF warns that 40% of global jobs are exposed to AI, and many will be displaced. Kristalina Georgieva, the IMF’s managing director, says nations must act urgently: “Countries must establish comprehensive social safety nets and offer retraining to protect livelihoods and curb inequality.” The U.S. is more exposed than most other countries as white collar workers make up approximately 60% of the workforce, and they are the most susceptible to AI job displacement.

We’ve done it before. The GI Bill rebuilt America’s middle class after World War II. The Marshall Plan rebuilt Europe. But today? We’re spending over $1 trillion a year on defense, while teachers’ pay for classroom supplies out of pocket and rural communities still lack broadband. That’s not national security. That’s national shortsightedness.

The new Plan mentions tax incentives for AI training and pilot retraining programs—but stops short of guaranteeing safety nets or offering bold ideas like a national AI dividend. When tens of millions are impacted by this technology, such programs won’t even act as a Band-Aid solution. Without systemic economic buffers, we risk repeating the mistakes of past industrial transitions.

Scaling AI Means Building for All

Let’s be clear: this isn’t about slowing down AI research or pulling back from global leadership. It’s about matching our ambition in the lab with ambition in the living room.

Imagine every American child with access to a personalized AI tutor. Imagine AI-powered diagnostics reaching every rural clinic. Imagine smarter infrastructure, decarbonized energy grids, and lifelong education delivered through immersive technologies.

Sadly, this vision is still missing in today’s official blueprint. There’s no national plan for scaling the services that are needed to ensure a smooth transition to a post-AGI future.

That’s the race worth winning.

And yet, other nations are outpacing us—not in flashy AI demos, but in deployment readiness. China, for example, has more than doubled U.S. electricity production and the gap is growing rapidly. It is investing heavily in AI integration across manufacturing, medicine, logistics, and education. Europe is pairing its AI investments with strong worker protections and digital rights. They may be behind on raw model performance today, but as open-source AI narrow the gap with leading frontier labs, the advantages of initial invention matter less each day.

In this race, adoption beats invention. Distribution beats dominance. And if America doesn’t invest in the systems to scale AI equitably and widely, we’ll find ourselves like the hare in the fable—fast out of the gate but napping at the wheel when it matters most. The faster we try to deploy when the social infrastructure isn’t ready, the more internal chaos we will create within our borders.

A New Definition of Strength

Real strength in the AI age won’t be measured by model size or benchmark scores. It will be measured by the resilience of our economy, the dignity of our workforce, and the trust we earn around the world.

We need a Universal Basic Infrastructure: Guaranteed access to high-speed internet, retraining programs, mental health support, and maybe even a national AI dividend—a 21st-century UBI that rewards care, creativity, and community contributions, not just code and capital.

This isn’t idealism. It’s realism. In Stockton, California, a UBI pilot gave low-income residents $500 a month. Within a year, full-time employment and entrepreneurship increased, while depression and anxiety dropped. It didn’t make people lazy. It gave them hope.

Now imagine that scaled across the nation. Hope, as policy. We are the most wealthy nation on the planet, yet we are behind most in protecting our population against looming change.

Lead With Vision, Not Fear

We don’t have to face the future alone or in fear. Cooperation is not weakness, it’s wisdom. The U.S. and China could co-invest in clean energy, drug discovery, robotics, or AI safety—creating the protocols and systems that protect the world. A global “CERN for AI” could align the brightest minds on a common mission: To make AGI safe, fair, and beneficial to all.

Unfortunately, the current Plan offers no vision for this. Despite recognizing national security risks from frontier models, it frames safety as a domestic surveillance challenge, not a shared global mission. The Plan’s adversarial framing reinforces a zero-sum mindset that undermines the very cooperation we need for safe AGI. The obsession with “dominance” risks triggering a spiraling AI arms race where AI could get out of control, and bad actors have more room to hide and strike.

Just as America led the world in building democratic institutions after WWII, we can now lead in building the global architecture of the AI age.

This Is the Moment

We are not deciding who leads in AI. We are deciding what kind of world we want to live in and what kind of future we’ll hand to our children and grandchildren.

Do we want to be the country that built the smartest machines but left its people behind?

Or do we want to be remembered as the nation that used AI not to dominate, but to elevate? That turned intelligence into abundance, not just for the few, but for the many.

We don’t need to win the wrong race. We need to build the right path forward.

It won’t matter who crossed the AGI finish line first. What matters is who’s still standing tall on the podium at the end of the century.