Congress Sounds the Alarm as US Companies Quietly Build on Chinese AI

A Security Warning Hidden in a Cost-Cutting Decision

Chinese AI models have gained traction among U.S. firms as they've closed the performance gap with American rivals while being cheaper to use. For many companies, it seems like a simple business decision. But Congress sees something far more dangerous lurking beneath the surface — and a growing investigation is now forcing some of America's most recognizable tech companies to answer for their choices.

The House Committee on Homeland Security and the House Select Committee on China are jointly investigating the rise of Chinese AI adoption among U.S. firms, an effort that began in April with letters to Anysphere and Airbnb. The probe specifically targets Airbnb and Anysphere over their use of Chinese-developed AI models, including Alibaba's Qwen and Moonshot AI's Kimi.

What These Companies Actually Did — and Why It Alarmed Washington

Anysphere's coding tool Cursor — which SpaceX is acquiring for $60 billion — built its Composer 2 model using Kimi, developed by Beijing-based Moonshot AI; the company declined to comment on the probe. In a separate letter to Airbnb, the committee chairs requested additional information regarding the company's reliance on an AI model developed by Alibaba for customer service operations, raising concerns about the company's preference for Chinese-developed technology. Airbnb had publicly stated it was relying on Qwen over American alternatives because it is "fast and cheap."

Airbnb said its AI activity ran overwhelmingly on U.S.-origin models, adding that a limited number of Chinese open-source models it used operated only through approved U.S.-based service providers. But lawmakers aren't satisfied with that explanation. Committee chairs argued that decisions to build products on Chinese Communist AI models threaten critical infrastructure Americans use every day, that the models are trained by China's censorship regime and introduce hidden vulnerabilities, and that Chinese AI companies are beholden to Chinese law and could turn over data they collect to the Chinese government if asked.

The Scale of the Problem Is Staggering

In late 2024, Chinese AI models reportedly accounted for an estimated one percent of global AI workloads. By the end of 2025, that share had reportedly grown to an estimated 30 percent. That explosive growth is exactly what's driving the urgency on Capitol Hill. The investigation follows an April 2026 memo from the White House Office of Science and Technology Policy warning that foreign entities, primarily based in China, are conducting deliberate, industrial-scale campaigns to distill U.S. frontier AI systems through proxy accounts and other coordinated methods.

Andrew Garbarino, chairman of the U.S. House Committee on Homeland Security, warned that "the Chinese Communist Party is no longer just nipping at our heels in artificial intelligence; it is racing to close the gap in some of the exact capabilities that will shape the future of cybersecurity," adding that reports of a Chinese open-weight model matching leading U.S. models in vulnerability discovery are "highly alarming." Meanwhile, the State Department said the "growing use of Chinese AI models by U.S. companies raises serious concerns," and that those models "are designed to advance Beijing's narratives, censor dissent, and reflect CCP ideology and values."

The Hard Question: What Can Washington Actually Do?

While some government departments have banned the usage of Chinese AI models including DeepSeek, adoption of them by U.S. companies is not prohibited. Tech chiefs, including crypto company Coinbase's Brian Armstrong and AI startup Lindy's Flo Crivello, have been publicly touting the use of models from China to reduce costs. That economic pull makes any legislative fix complicated.

Since Chinese open-source models are freely available online, an outright ban could run into First Amendment protections around speech, according to Kyle Chan of the Brookings Institution. The committees are also examining whether the United States has a sufficient open-weight AI strategy to ensure American companies and cyber defenders are not forced to choose between expensive or restricted U.S. models and cheap, capable PRC-developed alternatives. As Rep. Andy Ogles put it in June: "When the cheap, capable, easy option for an AI model is Chinese, the rest of the world will build on it." Until American alternatives become more competitive on price, that warning may prove difficult to heed — no matter how loudly Congress sounds the alarm.