By Doug Kelly
When President Trump takes the podium for tonight’s State of the Union Address, America’s lead over China in artificial intelligence (AI) will be thinner than most people realize. The two nations are locked in a fierce competition for global AI leadership, and the winner will shape the world’s economy, security, and digital future for decades to come.
The American Edge Project’s (AEP) Visual Guide to the U.S.–China AI Race makes the challenge clear: across the four pillars of AI strength – energy, compute, talent, and adoption – China already leads in three. America’s sole remaining edge is compute power: the data centers and advanced chips that fuel American AI. But that advantage is under pressure.
China’s Five-Part Plan to Overcome America’s AI Advantage in Compute
That pressure is not accidental. It is the result of a coordinated national strategy by China.
- First, China has made AI infrastructure a national mission. It treats AI infrastructure like railroads or highways: as strategic national assets built with urgency. Beijing invested $1.4 trillion into its tech sector this decade, with projections reaching $2.8 trillion by 2030. In 2024 alone, China’s AI computing capacity surged 74 percent. It has built the second-largest data center infrastructure in the world, controlling roughly one-fifth of global capacity, a dramatic rise from a decade ago. China recognizes that data centers are essential infrastructure in a serious global competition.
- Second, China is chipping away at America’s semiconductor advantage. For years, advanced chips were America’s firewall. That lead is narrowing. China’s most advanced domestic AI chips, such as Huawei’s Ascend 910B and 910C, are now powering large model training and delivering results comparable to leading U.S. chips while cutting costs by roughly 20 percent. By 2026, China is projected to have the world’s largest share of chip production capacity, building the industrial depth to support its AI ambitions. America’s chip edge is still significant, but it is no longer a permanent moat.
- Third, China is surging innovation through efficiency and scale. DeepSeek, built on less advanced hardware at a fraction of the cost of leading American models, showed the world that compute efficiency can substitute for compute superiority. That progress is not theoretical. Google DeepMind CEO Demis Hassabis recently said China AI models may only be “months” behind those of the United States. China has released more than 1,500 large AI models and now holds roughly 60 percent of global AI-related patents. Combining ingenuity with efficiencies, China is scaling real innovation.
- Fourth, China is exporting its full AI tech stack to lock in the developing world. Through its Digital Silk Road (DSR), China has deployed its AI systems and digital infrastructure in dozens of developing countries. This effort constructs an integrated ecosystem of chips, cloud services, data centers, and financing that, once embedded, is difficult to displace. In fact, one leading analyst warned last week that most of the world could be running on a “Chinese tech stack” within five to ten years. Every nation that adopts that stack is one more that moves further from American technology, American standards, and American values.
- Finally, China is backing all of this with the full power of the state to undercut American competitiveness. Beijing has maintained a comprehensive national AI strategy since 2017 and backs its AI companies with subsidized energy, multi-billion-dollar state investment funds, and government-directed procurement. In 2024, China generated twice as much electricity as the United States. Just one year later, it added eight times more new electric power than the U.S., underscoring Beijing’s commitment to making China the world’s leading AI power by 2030.
What America Must Do To Maintain AI Leadership
The picture is sobering, but this race is far from lost. America still designs the world’s most advanced chips and has the most innovative private sector on earth. We need a national strategy – a single playbook – that builds with urgency and efficiency by allowing every sector plays its strongest role. That means:
- The private sector should lead in building and deploying AI, innovating and competing against foreign adversaries while Washington clears the path.
- The federal government should codify and fund the nation’s AI Action Plan into durable law, establish light-touch national AI guardrails, and create a unified framework that prevents a 50-state patchwork of conflicting rules and red tape. Washington must also secure critical supply chains, and expand science, technology, engineering, and mathematics (STEM) education, and high-skilled immigration pipelines.
- State lawmakers should stop trying to regulate how AI models work and instead focus on American competitiveness: expanding energy generation and grid capacity; accelerating permitting for data centers, chip fabs, and grid interconnections; strengthening skilled trades workforce pipelines; and driving AI adoption across industries and state and local governments.
America’s Blueprint for Success: Building for Community and Country
We need to treat data centers for what they are: essential national infrastructure. Just as highways, power plants, and broadband networks form the backbone of our economy, data centers are now foundational to American security and competitiveness. The real question is not whether to build them, but how to build them well.
Yet in today’s political climate, building essential national infrastructure is too often cast as a zero-sum tradeoff of national strength versus local interests, big tech versus small towns. In reality, it’s about delivering true local gains today while strengthening the country for the long term.
That win-win requires partnership and balance. Communities deserve a voice. Developers must be responsible neighbors and long-term investors. Local and state leaders must weigh legitimate community concerns alongside economic opportunity and national interests with clear eyes. Local input matters. But so does getting critical infrastructure built.
Every community that hosts a data center strengthens America’s ability to compete with China while growing its tax base, creating quality jobs in construction and skilled trades, and attracting long-term investment and talent to the region.
America has led before in space, in computing, and in every defining technological era. We can lead in AI. But victory is not guaranteed, and we will not win divided. It’s time to build urgently and together.
###