By Doug Kelly 

China is going all in to win the race for artificial intelligence (AI): It has a national strategy and is investing hundreds of billions of dollars to carry it out.

The stakes are enormous. Whoever leads AI in will gain decades-long advantages in national security and economic competitiveness, along with heavy influence over the globe’s digital infrastructure and the free flow of information.

But while U.S. policymakers debate new AI regulations and data center bans, China is moving forward on three fronts at once: building its own AI infrastructure, working to slow ours, and exporting its AI models and authoritarian values around the world. Any one of these would demand a serious American response. Together, they demand one now.

Strategy One: China Is Building Its Own AI Infrastructure

China treats AI as a national mission, and its government is directing the build from the top down to be independent of any U.S.-made technology. Key elements include: 

  • Targeting a $1.4 trillion AI industry by 2030, including a $295 billion construction plan over five years to build a nationwide network of AI data centers.
  • Eliminating reliance on American technology. State-funded projects must now source at least 80 percent of their AI chips from domestic suppliers and away from American-made chips. Chinese labs are already training frontier models entirely on Huawei latest chip, proof they can build cutting-edge AI without American chips.
  • Deploying AI at industrial scale. China runs more than 30,000 AI optimized smart factories covering 80 percent of its manufacturing sectors. These factories generate shorter development cycles, higher productivity, and lower defect rates, giving China a major competitive advantage. What’s more, under Beijing’s military-civil fusion policy, advancing that same industrial base also arms the People’s Liberation Army.
  • They are stealing AI tech and planting code vulnerabilities. To train its own models, Chinese AI companies run massive “distillation attacks” against American AI companies, in one case using roughly 25,000 fake accounts generating 28.8 million conversations with a U.S. AI model to train its own. A 2026 Booz Allen Hamilton study found that Chinese coding tools produce measurably more vulnerable code when they detect a U.S. government user. These flaws look like ordinary code, but Booz Allen calls them a “sleeper agent” risk that could compromise data, trade secrets, and security.

Strategy Two: China Is Working to Slow America’s AI Build Out

Building fast is only part of China’s strategy. The other half is slowing down America’s AI efforts.

  • China is cutting off the materials we need. China controls nearly 80 to 90 percent of the refined rare earths needed for AI chips and hardware. In June 2026, Beijing imposed import restrictions on ten U.S. companies that America is counting on to build a domestic supply chain of these materials.
  • China is trying to turn Americans against the AI infrastructure we need. In July 2026, the New York Times reported on a study showing China, Russia and Iran have sought to use their state media outlets to turn the controversy over data centers in the United States into “a domestic fracture point.” Former Commerce Secretary Gina Raimondo warned that “much of the anti-data-center rhetoric online is propaganda funded by China,” adding, “Who would be the happiest person in the world if America stopped AI? President Xi.” Interior Secretary Doug Burgum similarly warned that projects are being “bombarded with foreign-directed propaganda to try to block these from being built,” and House lawmakers have asked the FBI to investigate possible Chinese influence.

Strategy Three: China Is Exporting Its AI Models and Its Values

China’s endgame isn’t just to build good AI. It’s to make the world run on Chinese AI.

  • The AI model gap is closing, and U.S. companies are already switching. China-based Z.ai’s newest model ranks just behind the top U.S. systems at a fifth of the cost, and U.S. firms now route more than 30 percent of their AI usage through Chinese models, up from 4.5 percent a year ago.
  • Chinese wants to make the world dependent on its technology. China’s AI models account for 41 percent of downloads, passing U.S. models outright. Through its Digital Silk Road, China is exporting its full AI stack across Africa, Latin America, and Southeast Asia. And it is trying to rewrite the global tech rules, proposing a new international AI body open to any government, with no democratic or human-rights standards, and designed to make authoritarian norms the global default.
  • Every Chinese model ships with Beijing’s politics built in. Chinese law requires models to reflect “Core Socialist Values,” and they refuse politically sensitive tasks at rates as high as 80 percent. Ask a Chinese model what happened in Tiananmen Square in 1989 and it refuses to answer, even while it readily explains American events like the Kent State shootings.

What Washington Must Do

AI is America’s modern-day moonshot. To win, we must match and exceed China on these same three fronts.

  • Build faster. Policymakers should codify and fund the American AI Action Plan, pause the patchwork of conflicting state AI laws, build the AI infrastructure necessary to win, especially data centers and the energy grid.
  • Protect our supply chains and our intellectual property. We must continue to reshore and friend-short to allied nations critical supply chain elements, scale up domestic rare-earth processing, treat industrial-scale model theft as the national-security threat it is, and keep untrusted Chinese AI models out of government and critical infrastructure before it is embedded, not after.
  • Aggressively compete for the world. Avoid self-defeating restrictions that keep American AI models out of the markets China is racing to capture, win the Global South through public-private partnerships, and lead on democratic AI standards instead of ceding that ground to Beijing.

It matters greatly which country builds the future. China is embracing a “whole of society” approach to winning the AI race. America must have that same level of commitment. The world is counting on us to get it right.