The competition between the United States and China over artificial intelligence is most often discussed in economic terms: who will build the better models, who will win the next round of investment, who will set the standards. Pat Condo, the CEO of Seekr Technologies, is framing it in different terms. According to a briefing on his remarks, Condo warns that the AI race with China is a national security risk and points to the sale of Nvidia H200 chips to Chinese firms as a specific focus of concern.

The warning stands out because of where it comes from. Most high-profile statements about AI and national security come from government agencies, military officials, or policy researchers. Condo runs a technology company and operates inside the commercial AI economy. A call from that vantage point carries practical weight that abstract policy debates do not. He is describing a dynamic he can observe from inside the industry.

The H200 chip sits at the center of that concern. It is one of Nvidia's processors for demanding AI computing workloads, the kind of hardware used to train and run large machine learning models. The briefing does not provide volume figures, buyer names, or information about whether the sales fall within existing export restrictions. The briefing makes clear that Condo sees the continued movement of high-end Nvidia hardware to Chinese companies as part of the security problem rather than an ordinary business matter.

That is a significant position for an industry figure to take. U.S. chipmakers have commercial reasons to want Chinese customers. China is one of the largest semiconductor markets in the world, and selling into it helps American companies amortize enormous research and development costs. By treating chip sales to China as a security risk, Condo is effectively arguing that the security calculation should outweigh the commercial one. That is not the default view inside the technology industry, and it is worth taking seriously precisely because it cuts against self-interest.

The underlying concern is the dual-use nature of AI. The same computing capacity that powers language models, search tools, and productivity software can also support military planning, surveillance systems, cyber operations, and autonomous defense work. Advanced chips are the raw foundation of that capacity. High-end hardware flowing to companies in a country that has made AI a strategic priority creates a real potential for those chips to serve military or intelligence purposes. That risk is the reason such hardware is so tightly watched in the first place.

Based on the briefing, Condo's warning does not include a specific prescription. He does not call for new export controls or an outright ban on H200 sales to China, and he does not claim that a specific misuse has already occurred. The framing is about risk, not about a completed threat. That distinction matters. "National security risk" is a broad category. It can mean an imminent danger, or it can mean a slow erosion of technological advantage that leaves the United States worse off a decade from now. Both readings are present in this kind of warning, and each implies a different policy response.

There is also a counterargument worth acknowledging. Restricting chip sales does not necessarily slow China's AI progress. Chinese companies have invested heavily in domestic chip development, and advanced hardware can cross borders through third countries regardless of who officially approves the sale. On that view, export restrictions mainly push revenue to Chinese competitors while failing to deny them the capability they want. Condo's warning tilts the other way. It treats the flow of advanced chips to Chinese firms as a serious risk in its own right, even if every sale cannot be traced or blocked.

The broader context is a race that extends far beyond any single product. The U.S.-China AI competition is being fought across model development, chip manufacturing, talent, and computing power. The H200 sales are a narrow but visible slice of that larger contest. They are a useful focal point because they are concrete. A chip sale is easier to debate than an abstract technological trajectory. The chip has a known maker, a known capability, and a known destination, which gives the debate a concrete subject.

For readers who follow this story, the significance of Condo's warning is that it comes from inside the industry. Government assessments of China's AI ambitions are abundant, but those assessments are produced by institutions whose job is to look for threats. A commercial CEO choosing to describe the AI race as a national security risk is a different kind of signal. It suggests the concern is visible to people who are not paid to be alarmed and who have no obvious incentive to dramatize the situation.

The open question is whether the warning changes anything. The debate over advanced chip exports to China is already one of the most contested policy battles in the technology world, with security hawks on one side and commercial interests on the other. Condo's intervention adds a prominent voice to the security side of that argument. Regulators, lawmakers, and Nvidia itself may or may not adjust course in response. The warning sharpens the terms of the conversation: the sale of powerful AI chips to Chinese firms is not merely a trade question. At least one U.S. AI CEO views it as a matter of national security, and he has said so.