U.S. lawmakers have demanded information from DoorDash regarding its use of Chinese artificial intelligence models. This scrutiny comes as American businesses increasingly adopt systems developed in China, driven by the relentless pursuit of lower operating costs.
The chairmen of two House committees, conducting a joint investigation into the security implications of U.S. companies using Chinese AI, asked DoorDash for documents on its evaluation and deployment of these systems. A DoorDash spokesperson stated the company "proudly supports American AI leadership" and aims to ensure AI benefits "Main Street," not just the largest corporations. This corporate rhetoric attempts to mask the underlying drive for profit maximization.
Capital's Pursuit of Profit
DoorDash founder Andy Fang publicly noted the company delegates "lower-level AI work" to Kimi K2.6, a Chinese model from Moonshot AI. DoorDash's AI research lab on X claimed Kimi K2.6 and Anthropic's Fable 5 "vastly outperform" other models, including Anthropic's Sonnet 4.6 and Opus 4.8, "at a cheaper cost." This starkly illustrates capital's primary motivation: reducing expenses to boost surplus extraction.
Other tech chiefs echo this sentiment. Coinbase's Brian Armstrong and AI startup Lindy's Flo Crivello have openly advocated using Chinese models to cut costs. The House committees themselves acknowledged that U.S. companies, from startups to large firms, deploy PRC-developed open-weight models for their "competitive capabilities, lower costs, greater customization and alternatives to reliance on a small number of proprietary model providers." These "practical considerations" are, in essence, the dictates of capital seeking maximum efficiency and profit.
The State's Role in Protecting Capital
The state, however, is not a neutral arbiter in this inter-capitalist competition. The House Committee on Homeland Security and the House Select Committee on the Chinese Communist Party lead the investigation, citing "national security concerns." Andrew Garbarino, chairman of the Homeland Security Committee, warned that the Chinese Communist Party is "racing to close the gap" in AI capabilities, particularly in cybersecurity. This framing of national security serves to justify state intervention aimed at protecting domestic capital interests against foreign competition.
While some government departments have banned Chinese AI models like DeepSeek, their adoption by U.S. companies remains largely unrestricted. President Donald Trump signed an AI executive order in early June, giving federal agencies 60 days to develop a key framework. This deadline, which passed on Saturday, August 1, drew intense attention from Silicon Valley's most powerful figures.
Tech leaders, including OpenAI CEO Sam Altman and Nvidia's Jensen Huang, met with lawmakers and Trump administration officials in Washington, D.C., to influence upcoming decisions. Trump's executive order, notably "thin on details," asks AI companies to voluntarily submit models for government evaluation. Treasury Secretary Scott Bessent, Defense Secretary Pete Hegseth, White House Chief of Staff Susie Wiles, and Commerce Secretary Howard Lutnick were tasked with defining the specifics. This process highlights the state's collaboration with dominant capital, not its regulation for public good.
Inter-Capitalist Contradictions
The rapid advancement of Chinese open-weight models, offering "cost-efficient alternatives" to the closed "frontier models" from OpenAI and Anthropic, sparked a fierce debate within the tech sector. These open-weight models allow companies to download, modify, and self-host the technology, further reducing reliance on proprietary providers.
In a rare display of unity, leading tech executives, including Huang, Elon Musk, Microsoft, Meta, and Palantir, opposed any potential ban on Chinese open-weight models. Huang's letter urged policymakers to avoid "premature restrictions," a clear defense of capital's freedom to pursue the most profitable options, regardless of origin. OpenAI eventually joined this stance, though Anthropic outlined its position separately.
The Trump administration's framework will involve a "classified benchmarking process" to assess models' cyber capabilities and encourage companies to work with the government to determine "trusted partners." White House Chief of Staff Susie Wiles, in a late June post on X, declared, "The government and private sector have worked together in a way we have never seen before and this foundation of America First is unprecedented." She asserted their "shared priority remains: get the best tech deployed as quickly and safely as possible," a statement that prioritizes the deployment of technology for capital accumulation, cloaked in nationalistic rhetoric.
Soon after the executive order, Anthropic temporarily disabled access to its Fable 5 and Mythos 5 models due to an export control directive. Two weeks later, OpenAI agreed to limit its GPT-5.6 rollout to "trusted partners" at the government's request. Both companies eventually rolled out their models more broadly, demonstrating the state's ultimate flexibility in accommodating the interests of its most powerful domestic tech monopolies, even as it feigns concern over foreign competition.