DoorDash is now at the center of a widening congressional investigation into whether American companies are becoming too dependent on artificial intelligence systems developed in China—a question that cuts to the heart of how the U.S. will compete in the technology race while protecting national security.
The House Committee on Homeland Security and the House Select Committee on the Chinese Communist Party sent a letter to the food delivery company requesting documents about its evaluation and deployment of Chinese AI models. The inquiry was triggered by a post from DoorDash founder Andy Fang acknowledging that the company delegates lower-level AI work to Kimi K2.6, a model developed by China's Moonshot AI.
DoorDash responded that it "proudly supports American AI leadership and is working to ensure AI benefits Main Street, not just the biggest companies," while pledging to engage with lawmakers about how it "safely and responsibly" uses both American-developed and open-weight models.
The Economics Driving the Dependence
The investigation reveals a structural problem: Chinese AI models are cheaper, faster to customize, and offer capabilities that rival American alternatives. DoorDash's AI research lab noted that Kimi K2.6 and Anthropic's Fable 5 "vastly outperform" other Anthropic models at "a cheaper cost." This isn't a matter of corporate recklessness. It's a rational economic choice in a market where smaller companies face cost pressures that giants like OpenAI and Anthropic don't.
Tech leaders including Coinbase's Brian Armstrong and AI startup Lindy's Flo Crivello have publicly defended their use of Chinese models to reduce costs. But the committees argue that practical business considerations don't eliminate national security risks. They wrote that U.S. companies—from large technology firms to startups—may adopt Chinese models "because they can provide competitive capabilities, lower costs, greater customization and alternatives to reliance on a small number of proprietary model providers." Those practical considerations, they cautioned, "do not eliminate the need for risk-based safeguards or diminish the national security concerns associated with growing dependence on models developed by entities subject to PRC jurisdiction."
Andrew Garbarino, chairman of the House Committee on Homeland Security, framed the stakes bluntly: "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." He highlighted that recent reporting showed a Chinese open-weight model can match leading U.S. models in vulnerability discovery and cybersecurity tasks—capabilities with direct national security implications.
A Divided Tech Industry, a Murky Government Response
The congressional pressure arrives as the tech industry itself is fractured over how to respond. In an almost unprecedented show of unity, the world's leading tech executives—including Nvidia's Jensen Huang, Elon Musk, Microsoft, Meta, and Palantir—signed a letter urging policymakers to avoid "premature restrictions" on Chinese open-weight models. OpenAI joined later. Anthropic did not, instead publishing a separate blog post outlining its position.
Meanwhile, the Trump administration is navigating its own regulatory framework. The president signed an AI executive order in early June, tasking federal agencies with developing a key framework within 60 days. That deadline arrived on August 1, drawing intense attention in Silicon Valley. The order asks AI companies to voluntarily submit their models to the government for evaluation ahead of public release, though the framework itself remains thin on details.
Sam Altman, OpenAI's CEO, met with White House Chief of Staff Susie Wiles this week, along with Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick. Altman told CNBC he'd reviewed a draft of the framework but didn't have "anything specifically in mind" as far as proposed changes. The administration has not officially taken a position on whether to restrict or ban Chinese models, though former AI czar David Sacks—who retains influence—has been vocal in opposing any ban.
Regulatory Whiplash and Unequal Burden
The regulatory uncertainty has already created complications for the major model developers. Anthropic had to disable access to its Fable 5 and Mythos 5 models shortly after the executive order to comply with Commerce Department export controls. Two weeks later, OpenAI agreed to limit the rollout of its GPT-5.6 model series to a group of trusted partners at the government's request. Both companies eventually rolled out their models more broadly, but the process revealed how the rules remain in flux.
White House Chief of Staff Susie Wiles wrote in late June that "the government and private sector have worked together in a way we have never seen before," framing it as an "America First" approach where the "shared priority remains: get the best tech deployed as quickly and safely as possible."
The committees have proposed that an effective federal approach should "scrutinize U.S. companies' reliance on People's Republic of China-developed models and strengthen the availability, security and competitiveness of American open-weight alternatives." This framing suggests the problem isn't simply competition—it's that American companies lack viable alternatives that don't require them to choose between cost-efficiency and national security.
Why This Matters:
The DoorDash investigation exposes a fundamental market failure in AI development. When Chinese models offer superior price-to-performance ratios, American companies rationally adopt them—but that rational choice, multiplied across the economy, creates concentrated dependence on systems controlled by foreign governments. The committees' recommendation to strengthen American open-weight alternatives acknowledges that regulation alone won't solve this problem; the U.S. needs to invest in competitive, domestically controlled options that don't force smaller companies to choose between their bottom line and national security. Without such investment, congressional scrutiny becomes a form of corporate punishment rather than a solution. The framework deadline suggests the Trump administration is moving toward a managed approach, but the tech industry's unified opposition to restrictions signals that any policy will face intense pressure from the most powerful companies in the world.