Congress Hears China Will “Buy or Steal” U.S. Tech to Win the AI Race, Underscoring a Growing Threat to America’s Security and Economic Future
The warning delivered on Capitol Hill this week was blunt, and Americans should not dismiss it as routine Washington rhetoric. At a congressional hearing, House Select Committee on China Chairman John Moolenaar said the pattern was clear: China depends on U.S. technology to keep advancing its artificial intelligence sector and is willing to “buy what they can, and steal what they cannot” to accelerate that ambition. Witnesses at the hearing argued that access to compute power remains the single most important ingredient in the AI contest and that China is aggressively trying to secure it through both legal and illegal channels. That message matters because the AI race is no longer just a commercial rivalry between companies. It is a contest tied directly to military capability, industrial leadership, cyber power, and long-term geopolitical leverage.
For the United States, the danger is not simply that China wants to build better chatbots or cheaper software tools. The deeper risk is that Beijing is trying to absorb enough of the American tech stack to narrow the strategic gap in a field that will shape everything from weapons systems and intelligence analysis to manufacturing productivity and critical infrastructure management. The congressional hearing, as reported by the South China Morning Post, emphasized that Chinese AI firms still rely heavily on American chips and American-origin components somewhere in the stack. That reliance is precisely why the issue is so urgent for U.S. policymakers: if China remains dependent on American technology while simultaneously finding ways around export controls, then U.S. innovation is not only powering American growth, it may also be helping a rival close the gap in capabilities Washington considers strategically sensitive.
This is not just a theoretical concern. Reuters has reported repeatedly that U.S. lawmakers are trying to tighten controls on the tools and chips that enable China’s AI and semiconductor progress. In February, Reuters reported that lawmakers pushed the State and Commerce Departments to further restrict China’s access to chipmaking tools, arguing that existing controls had not fully stopped Chinese technological gains. In April, Reuters also reported that lawmakers scaled back—but did not abandon—a bill targeting Chinese chipmaking, with the revised measure still preserving targeted export bans and continuing broader efforts to align allied controls in order to preserve U.S. leadership in AI. Those developments make clear that Washington’s concern is not episodic. It is part of an ongoing recognition that semiconductors, compute access, and AI infrastructure now sit at the center of national power.
Americans should also pay close attention to the smuggling problem, because it shows how difficult this competition has become to police. Reuters reported in 2025 that lawmakers introduced the Chip Security Act specifically to address reports that export-controlled U.S. AI chips were being smuggled into China. More recently, Reuters reported in February 2026 that the U.S. Commerce Department said China had not yet received Nvidia’s H200 chips, but also emphasized that preventing chip smuggling remained a top enforcement priority. That combination is revealing. It shows that Washington is not only trying to regulate legal sales but is also fighting a shadow market designed to bypass the rules entirely. When a strategic rival is willing to exploit smuggling routes, intermediary entities, and procurement workarounds to obtain advanced compute, this is no longer just a trade dispute. It becomes a matter of economic security and export-control credibility.
The hearing’s message that China will buy what it can and steal what it cannot also reflects a broader congressional view that the AI contest is inseparable from concerns about technology transfer and intellectual property leakage. While lawmakers often use strong language in hearings, that concern is rooted in a real pattern of U.S. anxiety over China’s access to advanced chips, semiconductor tools, and AI-enabling hardware. Reuters has covered repeated efforts by Congress to curb Chinese access to equipment that China still struggles to make domestically, including high-end lithography and related servicing. The logic is straightforward: if China cannot yet reproduce certain tools at the frontier, then preserving U.S. and allied control over those chokepoints is one of the most effective ways to slow China’s rise in strategic AI applications.
That matters for ordinary Americans because AI leadership is not an abstract trophy. It will influence the strength of the U.S. economy, the resilience of domestic manufacturing, the speed of scientific breakthroughs, the shape of future labor markets, and the balance of military power. If China can use American-origin chips, software ecosystems, and research exposure to rapidly narrow the gap, the consequences will not be confined to Silicon Valley boardrooms or think-tank panels. They could affect everything from the competitiveness of U.S. firms to the survivability of American networks and the sophistication of adversarial military systems. Congressional witnesses emphasized compute as the central strategic input for AI dominance for exactly this reason: whoever controls the hardware, scale, and data-processing capacity behind advanced models gains leverage across many sectors at once.
There is another layer to this challenge that deserves attention. Some voices at the hearing reportedly argued that U.S. immigration and research policies could themselves hinder American innovation if they are too clumsy or too restrictive. That is a real dilemma. The United States has historically led in technology because it attracts talent, funds open research, and commercializes innovation more effectively than most rivals. But openness becomes a vulnerability if the same system can be exploited to siphon know-how, circumvent export controls, or accelerate the ambitions of strategic competitors. The hard part for Washington is not deciding whether to be open or closed in absolute terms. It is building a system that remains innovative and attractive while becoming much harder to exploit. The hearing captured that tension well: America cannot afford either reckless openness or self-defeating overreaction.
The broader U.S.-China policy environment makes this even more urgent. Reuters reported in February that Democratic lawmakers criticized the Trump administration for shelving some China tech measures ahead of high-level diplomacy, saying those delays threatened national security. Even where lawmakers disagree over exact tactics, the underlying concern is bipartisan and persistent: U.S. policy cannot allow short-term political calculations or industry pressure to hollow out the controls meant to protect strategic advantages. That is especially true in AI, where time matters. A few years of unchecked access to American compute, tools, or research pathways could produce long-term consequences that are difficult to reverse once China’s domestic ecosystem matures further.
It is also worth noting that this anxiety is not just about one company like Nvidia or one Chinese firm like DeepSeek. Reuters has reported that evidence continues to emerge showing how U.S.-designed or U.S.-controlled technologies can be instrumental in Chinese AI progress even under restrictions. Earlier reporting tied Nvidia’s H20 chips to Chinese AI development, and lawmakers have repeatedly pointed to the role of American-origin compute in enabling Chinese systems that surprised Washington with their speed of progress. The concern is therefore structural. If Chinese firms can keep obtaining enough U.S.-linked hardware, directly or indirectly, then export controls risk becoming symbolic rather than decisive.
What should worry Americans most is that this is exactly the kind of challenge that can be underestimated until it becomes much harder to manage. AI competition does not usually arrive in the form of a dramatic headline announcing that America has suddenly “lost.” Instead, it shows up as a series of small erosions: more chip leakage than expected, more research spillover than intended, more policy compromises than advertised, more Chinese performance gains built on systems Washington thought it had constrained. By the time that cumulative drift becomes obvious, the strategic gap may already have narrowed far more than policymakers intended. The hearing’s language sounded sharp because some lawmakers believe that is precisely what is happening now.
The United States does not need to panic, but it does need to stay clear-eyed. China’s AI ambitions are public, well-funded, and tied to a national strategy that reaches well beyond civilian technology branding. American lawmakers are increasingly treating compute, chips, and research integrity as strategic assets rather than ordinary commercial goods. That shift is justified. If Beijing is indeed willing to buy what it can and steal what it cannot, then the United States is facing a rival that views American openness not only as something to admire, but also as something to exploit. The lesson for Americans is not that innovation should slow down. It is that innovation without protection will eventually help power the rise of the very competitor Washington is trying to contain.
