The United States still has the companies, talent and capital to remain the world’s dominant artificial-intelligence power. But the advantage is shrinking, and experts warn that America needs a national strategy that reaches far beyond the technology industry.
America’s lead in artificial intelligence is no longer guaranteed. While U.S. companies still dominate frontier model development, chip design, cloud infrastructure and venture investment, the global race is tightening as China accelerates deployment, expands open-source alternatives and turns AI into an industrial-policy priority.
A Wall Street Journal essay by Jim VandeHei argues that the window for American dominance “hasn’t closed,” but is narrowing fast. The central warning is that the U.S. is treating AI as a series of disconnected business, education and regulatory debates rather than as a generational national challenge.
The concern is not that America has already lost. In many areas, it remains ahead. U.S. firms such as OpenAI, Anthropic, Google, Microsoft, Meta, Amazon and Nvidia continue to shape the frontier of AI systems and infrastructure. Recent analysis also points to the depth of American private investment, with the U.S. attracting far more AI capital than any rival economy.
But China is competing differently. Rather than only trying to beat the United States at the frontier model layer, Beijing has pushed rapid deployment across manufacturing, logistics, urban systems and consumer technology. Chinese firms have also gained attention for cheaper, widely accessible AI models that challenge assumptions about how expensive advanced systems must be.
That creates a strategic dilemma for Washington. If America focuses only on building the most powerful models, it may miss the broader race to apply AI across the economy, government, schools and national security. Winning the AI race will depend not only on technical breakthroughs, but also on whether the country can adapt institutions fast enough.
Seven ideas now stand out.
First, the U.S. needs a real national AI infrastructure plan. Data centers, power grids, advanced chips and cloud capacity are becoming as strategically important as highways, ports and oil reserves once were. The White House’s AI Action Plan has framed AI dominance as a national-security imperative, but implementation will depend on whether the federal government can align energy permitting, chip supply chains and compute access at scale.
Second, America must modernize education quickly. AI literacy can no longer be treated as an elective skill for engineers. Schools, universities and workforce programs need to teach students how to use AI tools, evaluate outputs, protect privacy and work alongside automated systems. A country that leads in AI research but leaves most workers unprepared will struggle to turn innovation into broad prosperity.
Third, Washington should expand access to computing power beyond Big Tech. Frontier AI is increasingly expensive, giving the largest firms a structural advantage. Public research institutions, startups and universities need affordable access to compute so that innovation does not become trapped inside a handful of companies. Proposals such as a National AI Research Resource reflect this concern and aim to widen access to datasets, testing environments and computing capacity.
Fourth, the U.S. needs a faster path for high-skilled AI talent. America’s strength has always depended partly on attracting the world’s best scientists, engineers and entrepreneurs. If immigration bottlenecks push foreign graduates and researchers toward rival ecosystems, the U.S. will weaken one of its most important advantages.
Fifth, the country must prepare for labor-market disruption before it becomes a crisis. AI is already reshaping white-collar work, and policy remains behind the technology. Economists and policy researchers have warned that Washington lacks detailed plans for taxation, safety-net reform and worker transition if AI rapidly changes employment patterns.
Sixth, the U.S. should defend critical AI supply chains without isolating itself from allies. Export controls on advanced chips remain one of Washington’s most powerful tools, but controls work best when paired with allied coordination, domestic manufacturing and clear rules that protect innovation while limiting strategic leakage to adversaries.
Seventh, America needs a governance model that builds public trust. A backlash against AI could slow deployment just as much as foreign competition. Concerns over job losses, misinformation, privacy, bias, water use and energy demand are real. If policymakers ignore them, local opposition to data centers, schools’ suspicion of AI tools and workers’ fear of automation could harden into a national resistance movement.
The race, then, is not simply Washington versus Beijing or OpenAI versus Chinese rivals. It is a test of whether American society can organize itself around a technology that may reshape economic power, military capability and everyday work.
The U.S. still has enormous advantages: deep capital markets, elite universities, leading semiconductor firms, a culture of entrepreneurship and the world’s strongest AI labs. But advantages can erode when policy is slow, infrastructure is blocked, education lags and public trust collapses.
America does not need panic. It needs coordination. The next phase of the AI race will reward countries that can build, deploy, regulate and educate at the same time.
The window remains open. But it is no longer wide.
