Artificial intelligence (AI) leadership is essential to U.S. national security, global stability, and economic strength – and urgent actions are required to ensure America prevails in the defining competition of the 21st century. As President Trump’s AI Action Plan notes, “Whoever has the largest AI ecosystem will set global AI standards and reap broad economic and military benefits.”
Today, putting America First means ensuring America is first in AI. President Trump’s AI Action Plan is a bold start. His plan addresses the infrastructure, regulatory, talent, and innovation challenges necessary for America to lead in AI. But the AI race is a marathon, not a sprint. Future American success depends not only on the plan’s rapid execution but also on a sustained national commitment to innovation and global adoption of American AI.
President Trump’s AI Action Plan recognizes that AI leadership is not simply a matter of funding research but of building a whole-of-nation ecosystem hospitable to rapid experimentation and deployment. Yet it is only the beginning. American success will require rapid execution, close partnership with Congress, and sustained bipartisan commitment to unleashing American innovation over the coming decade.
Leading in AI will require two distinct but simultaneous efforts: innovation and global adoption. It is not enough to drive the development of leading-edge AI models, cloud computing infrastructure including chips, and software platforms; we must also support the adoption and integration of American AI throughout the wider free world. The competition will turn largely on which power is able to innovate and drive global adoption fastest.
The AI race is America’s to lose. Attempts to slow China or preserve America’s lead through heavy-handed interventions in one of history’s fastest-moving markets are likely to prove particularly counterproductive. Success in the global AI competition does not depend on whether China can be slowed down at the margins of cutting-edge technologies, but rather on whether America can continue to produce revolutionary breakthroughs in innovation and adoption. What is needed is a national commitment to fostering American success through low regulation, competitive tax policies, strong trade partnerships, and an environment that fosters innovation.
I. AI and the New Superpower Competition
AI is a key lever of geopolitical power and could quickly eclipse prior technological revolutions, such as nuclear weapons, cyberspace, and even electricity. Every major dimension of national power – economic dynamism, military strength, intelligence capabilities, and even diplomatic influence – is being reshaped by AI. Countries that harness AI to enhance productivity, accelerate innovation, and modernize their security establishments will enjoy outsized advantages; those that fall behind will find themselves dependent on foreign technology and vulnerable to coercion.
In previous eras, breakthroughs in nuclear weapons, precision-guided munitions, and cyberspace created temporary windows of advantage that could be exploited by a handful of states. AI, by contrast, is an inherently general-purpose technology. It can be embedded into every sector, from finance to manufacturing, from logistics to health care, and from intelligence analysis to space operations. That breadth makes AI unique as a foundation of national power—and raises the stakes of the current AI race between the United States and the Chinese Communist Party (CCP).
To ensure the success of American AI, the United States must win the race for global adoption. That means winning not only on performance, but also on quantity and price. It means building an AI stack—from chips and cloud computing infrastructure to models and applications—that is so competitive, scalable, and trusted, that governments and firms around the world choose it over Chinese alternatives.
Unfortunately, China is quickly closing the gap in every technological domain relevant to AI. China’s strategy of commodifying the AI supply chain and offering high-performance Large Language Models (LLMs) for free demonstrates a national commitment to winning the competition for global adoption.
In rising to this challenge, America leads with its incredible private sector, the world’s greatest ecosystem of innovation. That ecosystem is America’s competitive advantage in the AI competition. It allows the most futuristic and revolutionary ideas to be developed, commercialized, and scaled around the world with amazing speed.
But China has several competitive advantages in the AI competition: it is able to achieve enormous economies of scale and maintains a great advantage in production costs. Its people are incredibly hard-working, increasingly confident, and hungry for success. They have developed a huge talent pipeline based on widespread science, technology, engineering and mathematics (STEM) education, with half the world’s AI engineers. And while they suffer from too much control by an unaccountable totalitarian state, their industries are paradoxically very lightly regulated, significantly enhancing their advantage in cost and speed.
Meanwhile, America faces significant obstacles of its own in the competition for global tech leadership, chief among which are overregulation and fiscal disfunction. Unlike China, the United States has made little headway in tackling its own barriers.
America and China each have advantages and disadvantages in the new AI competition. The outcome will turn on which power is best able to capitalize on its advantages and overcome its disadvantages.
One key question is: Which power will prove better able to overcome the obstacles it faces?
II. National Power Depends on Winning the AI Race
The stakes in the AI competition could not be higher. AI is rapidly becoming an essential part of the technological-industrial base on which both economic and military power rest. As such, AI holds enormous promise for the free world.
An American-led AI revolution would dramatically increase labor productivity, quality of life, and national wealth. AI will empower American workers to focus on more creative, higher-value tasks, while algorithms handle routine analysis, pattern recognition, and optimization. Across sectors, AI will help firms reduce costs, improve quality, and compress development timelines. Over time, these productivity gains could unlock trillions of dollars in new economic value, strengthen the dollar’s role in the global financial system, and reinforce America’s position at the center of the world economy.
AI and robotics will help unleash an industrial renaissance. In advanced manufacturing, AI systems can predict equipment failures, optimize supply chains, and enable hyper-precise quality control. Robots will revolutionize every aspect of the economy. For example, empowered by AI computer vision and decision-making, robots can handle dangerous, repetitive, or highly detailed tasks that humans cannot perform nearly as efficiently at scale. These capabilities can make it economical to onshore production that once seemed permanently lost to low-wage competitors, thereby tightening the link between American innovation and American jobs.
The national security implications are even more profound. AI enhances battlefield awareness, speeds decision-making, and empowers autonomous and semi-autonomous systems across every domain—land, sea, air, space, and cyberspace. Commanders who can fuse data from sensors, satellites, and open sources into a coherent picture faster than adversaries will enjoy a decisive advantage. Precision targeting, air and missile defense, logistics planning, and electronic warfare are all being transformed by AI-enabled systems—and the side that leads in those applications will shape the future of warfare.
Intelligence agencies rely on advanced AI models to analyze vast datasets and detect emerging threats. Properly designed and governed, commercial and custom-built AI systems can sift through signals intelligence, imagery, financial data, and open-source information to spot patterns that humans would miss. But this requires that U.S. intelligence agencies and law enforcement have access to the best models and the broadest permissible training data. If American models are hobbled by regulatory constraints while foreign adversaries freely train theirs, we risk creating dangerous blind spots in our ability to detect and prevent future threats.
AI’s military and defense applications alone are enough to warrant a massive national awakening, not to mention investment. Maintaining the lead in AI is now critical to deterring, and if necessary, winning war. The ways in which war is waged are rapidly changing as new capabilities, applications, and doctrine emerge thanks to AI and novel technologies. AI is already being used in faraway battlefields to improve situational awareness, speed decision-making, and empower autonomous and semi-autonomous systems. Its criticality to modern battlefields will only increase in the coming years.
Take, for instance, two weapons systems at opposite ends of the cost continuum. First, with the help of AI, hypersonic missiles will become more accurate, better connected, faster, and possess longer ranges, making their combat potential far more deadly. These high-cost weapons in the hands of adversaries like China have the potential to inflict devastating damage on assets ranging from aircraft carriers to air bases in the Indo-Pacific as our defenses have yet to catch up.
Of potentially greater concern though, are the weapons at the opposite end of the cost spectrum: easy to make and mass-produceable unmanned aerial vehicles (UAVs). China is urgently developing autonomous systems and producing them at scale, thus rapidly improving its drone capabilities. But this technology has also led to an expanded range of threats and threat actors, as it is no longer just sophisticated state actors who can acquire and use this technology. This has forced the Department of War to invest significant resources into counter-unmanned aerial systems (C-UAS) technologies, but our forces remain unacceptably vulnerable to drone attacks.
Intelligence agencies, too, rely on customized commercial AI models to analyze enormous datasets and detect emerging threats. As important as it is to rebuild our military with new and advanced hardware, we must also ensure that the commercial AI models supporting our security agencies are trained on the best and most comprehensive datasets possible—so there are no blind spots in detecting threats.
For these reasons and more, both national security and economic power depend upon a successful, globally competitive “AI stack.” That is why President Trump’s AI Action Plan articulates a bold vision of global AI dominance. “Whoever has the largest AI ecosystem will set global AI standards and reap broad economic and military benefits,” says the AI Action Plan. While it is too early to tell who will have the largest AI ecosystem in the 21st century, we already know that whoever it is will have a significant presence in the Chinese market.
III. America Still Dominates but China is Poised to Take the Lead
For now, the United States still dominates the global AI landscape. America’s advantages include world-leading semiconductor firms, cutting-edge graphics processing unit (GPU) manufacturers, hyperscale cloud providers, research universities, and a vibrant ecosystem of startups. Our open innovation model, deep capital markets, and culture of entrepreneurship have enabled U.S. firms to pioneer frontier models and platforms that are far ahead of most global competitors. Allied democracies—from Japan and South Korea to the United Kingdom and Israel—are also natural partners in building a shared AI ecosystem grounded in the rule of law and market principles.
Yet Chinese development of AI technology is progressing at astounding speed in every domain, from semiconductors to GPUs, scalable AI rack solutions, and frontier models. Chinese firms are rapidly improving the quality of their models and producing sufficient quantity to substitute for American technology. Within a few years, Beijing could field an export-competitive AI stack that it can offer to developing countries as part of its broader Digital Silk Road strategy, binding them into Chinese technical standards and governance norms.
The great disadvantage China faces in the competition is how far behind the United States they remain in technology, startup ecosystem, depth of domestic financial market, and rule of law. China’s economy and society show hallmarks of too much state control: deflationary pressure, rising unemployment, discontent with excessive surveillance and oppression, and an aging political elite facing a fraught period of succession. Progress on rule of law issues, such as protection of intellectual property, still lags far behind progress in technology, which rewards imitation but punishes innovation. But in general, China is overcoming the obstacles it faces with increasing speed and ingenuity, and this is particularly true with respect to those obstacles most relevant to the AI competition: engineering and manufacturing.
China’s accelerating technological progresses is making it an increasingly formidable and dangerous adversary. The CCP is integrating AI into military doctrine, authoritarian governance, and global tech exports. Domestically, AI is being used to power a surveillance state that tracks citizens’ movements, communications, and online behavior. Abroad, China is using AI tools to conduct cyberattacks, such as the Volt Typhoon campaign, suppress dissent, and influence foreign nations through digital infrastructure, media manipulation, and AI-generated disinformation. The CCP is racing to dominate in AI because to do so will enable China to control how AI will be developed and employed globally, and that in turn will maximize China’s influence and leverage over other states.
It is critical to recognize that the CCP views winning the AI race as intrinsically connected to their quest for national rejuvenation and great power status. Just as the results of the previous industrial revolutions changed or solidified who possessed hegemonic influence, the CCP believes that if it wins this fourth industrial revolution and dictates how the world develops and employs AI, China will then be primed to supplant the United States as the leading power.
With this end goal in mind, Beijing is directing private-sector innovation from the top down. They are also using its Digital Silk Road initiative and similar programs to proliferate Chinese standards and Chinese technology into the Global South and thus winning significant global market share of AI infrastructure. Their open weight models are gaining significant traction globally, challenging American models on cost and accessibility if not yet on capability. The United States must counter these efforts with urgency.
There is clear evidence that China is employing AI to bolster its security apparatus domestically and pursue its revisionist aims abroad. If left unchecked, Beijing will accelerate its use of AI to conduct devastating cyberattacks, suppress dissent, engage in corporate espionage, and influence foreign nations and elections through digital dirty tricks. A delayed or divided American response to this threat would invite long-term strategic disadvantages and global instability, irreparably harming American interests.
If Beijing succeeds in exporting its AI stack and governance model, the world could drift toward a future in which key networks, platforms, and standards are controlled from Beijing rather than shaped by free societies. The window for the United States to build the foundations for a democratic, rules-based AI order is measured in years, not decades. Failure to act decisively now would be a historic mistake.
IV. Obstacles to American Dominance in the AI Revolution
Most Americans alive today have only known a world in which the United States held clear military superiority, but the gap is closing fast, with China already matching or even exceeding American in areas, such as hypersonic weapons and fleet size. Preserving America’s edge will depend heavily on leadership in AI, as future conflicts are likely to favor the force that can process information, decide, and act with unmatched speed.
Overregulation is a major competitive disadvantage for American AI, and for American industry in general. Everything takes longer and costs more to build in the United States than it should, from data centers and transmission lines to semiconductor fabs and advanced manufacturing facilities. If policymakers layer heavy-handed AI-specific regulation on top of an already burdensome regulatory system, they risk suffocating the very innovation they hope to foment. Overregulating model training, fracturing policy across states, or limiting access to high-quality data weakens U.S. tools and strengthens adversaries.
The great obstacle America faces in the new tech competition is overregulation and fiscal disfunction: too much spending, taxation, and borrowing fueling an unsustainable national debt. And by contrast to China, the United States is making little progress in overcoming the obstacles it faces. Exacerbating the problem is a recent tendency toward performative anti-China bluster that seems less intent on assuring American victory than on desperately throwing obstacles in China’s path.
Preserving America’s technological edge requires winning global market share through trade. Understandably, export controls designed for dual-use items, such as the Export Control Reform Act (ECRA), have become an important part of the debate. But there has been a marked tendency to exaggerate the potential benefits and discount the potential downsides of export controls on AI. The best example was Biden’s AI Diffusion Framework, which President Trump wisely rescinded in May 2025. Worse, as China develops competitive substitutes for U.S. technology, export controls will hurt China less and America more. Indeed, export controls on general purpose AI technology may already have outlived their usefulness. For this reason, it is especially important to keep any AI export controls under presidential control, where they can be calibrated as circumstances warrant.
Cutting edge technologies, such as AI, are most successful when the private sector is empowered to make long-term, strategic plans for its research and development. If American companies are forced to constantly redevelop their compliance model, it will paralyze decision making, investments will dwindle, and innovation will be stifled. This is the self-inflicted wound the CCP is hoping for to push American companies out of global markets with cheaper and comparable offers through centrally directed and heavily state- subsidized industry.
Let there be no doubt, regulations will be important to ensure American AI is developed and adopted in accordance with our democratic values, but it should not come at the expense of innovation, which is the American economy’s greatest strength. Doing so will only enable China to take the lead and export its autocratic AI models globally.
Some in Congress and in foreign capitals are pushing to limit the training data available to AI models by rewriting long-standing fair use copyright laws. Such restrictions could have dangerous and unforeseen consequences, especially for national security applications. If American models cannot train on broad, representative datasets while foreign adversaries face no such constraints, the result will be weaker tools in the hands of our defenders and stronger ones in the hands of our rivals. Well-intentioned but overbroad rules could create precisely the asymmetry Beijing hopes for.
AI development also requires massive and reliable energy, advanced compute facilities, and a highly trained technical workforce – but America’s electricity supply is already dangerously constrained. Building AI infrastructure at scale requires significant new electricity generation and streamlined grid interconnection—yet modern hyperscale data centers are among the most energy- and water-efficient computing facilities ever built. The bottleneck is not demand but permitting: outdated processes delay the construction of generation, transmission, and facilities that would bring jobs and tax revenue to communities across America. Permitting reform is essential to fast-track data center construction, energy projects, grid modernization, and rare earth mining, yet Congress has so far only tinkered at the margins. Cloud providers have committed billions to energy investment and advanced cooling technologies that minimize water use, but these investments cannot deliver results if projects are stuck in multi-year permitting queues. If we cannot rapidly build out both clean and reliable baseload power, the AI revolution will stall at home even as it accelerates abroad.
The United States is also badly losing the talent war to China and other competitors. It’s difficult for the world’s brightest minds and top AI researchers, engineers, and entrepreneurs to study, work, and stay in the United States. At the same time, our K–12 and higher education systems are not producing enough homegrown technical talent to meet surging demand. Unless we fix these bottlenecks, American firms will be forced to move more research and development elsewhere.
The United States must view these physical and human assets not as routine domestic policy issues, but as pillars of national strength in a global AI contest. Data centers, grid capacity, transmission lines, ports, and advanced manufacturing clusters are all elements of national power in the AI era – so is immigration policy, workforce training, and research funding. If we continue to treat them as isolated debates rather than components of a coherent national strategy, we will struggle to convert our technological advantages into durable geopolitical leverage.
V. A National Strategy for Winning the AI Race
The federal government has a critical role to play in mitigating the risks of AI diffusion, building out necessary infrastructure, ensuring that America remains a beacon for the world’s most talented engineers and entrepreneurs, and – most important – creating a regulatory climate that fosters rather than stifles innovation and competition.
A pro-innovation approach requires a light regulatory touch in antitrust and export policy, and a sweeping “big bang” deregulation of American industry more broadly. Congress must codify the AI Action Plan to ensure it becomes durable national policy, not a passing executive initiative that can be reversed with the next change in administration. Modernizing antitrust enforcement for the age of AI means focusing on genuine threats to competition – such as exclusionary conduct or anticompetitive mergers – without punishing scale and success that are necessary for training cutting-edge models and building globally competitive platforms.
Export controls should be used only where specific national security risks clearly outweigh any impact on export competitiveness. Poorly designed controls that are too broad or too static will merely encourage foreign firms to develop alternative supply chains, eroding U.S. influence without significantly slowing adversaries’ progress. Instead, export policy should be narrowly tailored to key chokepoints in the AI stack, regularly updated in light of technological developments, and coordinated with allies wherever possible. The goal is to deny adversaries capabilities that would meaningfully threaten U.S. security while preserving the vitality and global reach of the American technology sector.
At the domestic level, a five-to-10-year pause on conflicting state AI regulations is essential to avoid fragmentation and paralysis. If every state adopts its own AI regime, companies will spend more time navigating 50 different compliance systems than building useful products and services. Congress should establish a clear federal framework that preempts conflicting state rules while preserving room for experimentation in areas that do not directly affect interstate commerce or national security. Within that framework, maintaining fair use protections will allow U.S. models to train on broad, high-quality datasets critical for defense, health, small business, and science.
Energy, infrastructure, and talent must be treated as core pillars of AI strategy, not afterthoughts. Regulatory reforms should fast-track data center construction, energy projects, and grid modernization, while preserving appropriate environmental safeguards. Policymakers should incentivize private investment in new generation and transmission through predictable permitting timelines and streamlined reviews. In parallel, Congress should enact immigration reforms that make it easier for top STEM graduates to remain in the United States, expand high-skilled visa programs tied to national security needs, and support apprenticeship and retraining pipelines that prepare American workers for AI-augmented jobs.
Finally, the United States must anchor its AI strategy in a broader alliance of free and open societies. Just as the Space Race galvanized a generation and led to decades of technological spillovers, American leadership in AI can underpin a new era of prosperity, security, and shared democratic values. Washington should deepen cooperation with allies on semiconductor supply chains, AI safety research, export controls, and standards-setting at international bodies. By working with partners to build a trusted, interoperable AI ecosystem, we can ensure that the digital infrastructure of the 21st century reflects principles of transparency, accountability, and human dignity.
Two versions of the digital future are emerging. In one, American and allied innovation drives an AI revolution that boosts productivity, strengthens deterrence, and extends the reach of freedom. In the other, an ascendant China exports an authoritarian AI model that normalizes surveillance, censorship, and coercive state power. The choice is still ours to make—but the window is closing. America can win the AI race by uniting behind a grand national strategy that supports AI development and adoption, unleashes our industrial and scientific strengths, and rallies the free world around a shared vision for the AI century.