US Congress debates whether to share or restrict AI technology exports to strategic competitors, with positions divided on export control strategy.

China recently achieved a symbolic but significant milestone by unveiling LineShine, the fastest supercomputer in human history. The United States had previously held that title through El Capitan at Lawrence Livermore National Laboratory in California. This shift raises a critical question: how did China manage this feat despite years of increasingly stringent U.S. export restrictions?

The U.S. imposes significant restrictions on technology exports to China, including graphics processing units and the machines used to manufacture them—components essential to supercomputer functionality. Yet LineShine does not rely on graphics processing units. China effectively circumvented American trade barriers by constructing a massive network of standard microprocessors instead.

This development illustrates a fundamental lesson: Washington's strategy of restricting Chinese access to U.S.-made semiconductors and related products has not achieved its intended effect. Rather than being halted, China has simply adapted and developed alternatives. The same pattern emerged earlier this year when the federal government reversed course on whether Nvidia could export its H20 chips for use in sophisticated Chinese AI models. The Trump administration correctly recognized that if China would develop AI regardless, it should use American chips. Yet China declined the offer, having already developed homegrown alternatives in response to the restrictions.

China's strategy is clear: work around Washington's restrictions by strengthening its own domestic technology sector, focusing on scale and establishing itself as a global leader in technologies the U.S. attempts to prevent it from importing or manufacturing. This is a smart approach, and America should learn from it.

Export restrictions certainly have a place in technological competition, particularly given that China is widely considered a national security threat. However, restrictions alone cannot secure American sovereignty or economic future. If Washington's answer is merely to classify more technology as restricted and isolate it from global markets, America risks losing the very competition it seeks to win.

This concern applies especially to proposals like the Remote Access Security Act. The bill goes far beyond preventing sensitive technologies and high-risk applications from reaching China. Instead, it would prohibit U.S. providers from offering services to any companies that employ Chinese nationals, regardless of how benign their technology or whether their customers are based in allied countries. This extraordinarily broad approach risks shrinking American technological reach rather than expanding it, making U.S. platforms less attractive to international customers.

Those customers will not stop using AI simply because Washington makes American services harder to access. They will migrate elsewhere. Every customer Washington drives away from American cloud providers is a customer Huawei need not win on merit. Through such policy mistakes, the U.S. would hand Chinese technology companies new business—not containing China's influence but helping it expand.

The real objective should not simply be denying American technology to China. It should be ensuring that the rest of the world continues building on American technology rather than Chinese alternatives. Winning the AI race requires more than denial; it demands dominance in adoption.

Just as China strengthens its domestic technology manufacturing, the U.S. should do the same. The foundation exists. Though the U.S. invented the semiconductor only to see innovation and production shift to East Asia, the CHIPS Act of 2022 has begun reversing this trend by treating semiconductors as a matter of industrial policy and reshoring production. A Taiwanese company is building a $165 billion semiconductor manufacturing campus in the Arizona desert—one of the most significant construction projects globally—demonstrating the potential for domestic growth among U.S. companies.

The United States retains deep advantages over Chinese technology: its ability to attract talent, its formidable venture capital base, and an innovation ecosystem that is far more privatized than China's state-controlled system, allowing capital to chase new ideas more efficiently. These advantages simply require better leverage.

Owning the world's fastest supercomputer will not determine the winner of the AI race. Becoming the platform upon which the world depends will. Washington should spend less time erecting barriers around American innovation and more time ensuring the world continues choosing it over China.