x/theus · 
August 3, 2026 — DoorDash, America's largest food delivery platform, is now at the center of a high-stakes Congressional investigation after co-founder Andy Fang publicly revealed the company had integrated a Chinese AI model into its internal code review system. The case exposes the growing tension between cost-driven AI adoption and escalating national security concerns in Washington.
In early July 2026, DoorDash co-founder and CTO Andy Fang posted on X that the company had incorporated Kimi K2.6, an open-weight AI model developed by Beijing-based Moonshot AI, into its internal AI-assisted code review process. The setup routes lower-level development tasks to Kimi while reserving harder problems for Anthropic's Fable 5, an American frontier model.
Fang's rationale was straightforward: the hybrid approach delivered "better quality, cheaper cost" than using American models alone. DoorDash's AI research lab backed this up on X, stating that the Kimi K2.6 + Anthropic Fable 5 combination "vastly outperformed" the previous stack of Anthropic's Sonnet 4.6 and Opus 4.8 — at a fraction of the price.
The numbers bear this out. Kimi K2.6's API runs at roughly $0.60 per million input tokens, compared to $5.00 per million for Anthropic's Claude Opus 4.7 — a cost reduction of approximately 88%. For a company processing thousands of code reviews daily, those savings are substantial.
On July 31, 2026, the chairmen of two powerful House committees — John Moolenaar (R-MI) of the House Select Committee on China and Andrew Garbarino (R-NY) of the House Homeland Security Committee — sent a joint letter to DoorDash co-founder and CEO Tony Xu.
The letter sets two hard deadlines:
The lawmakers acknowledged the commercial logic — Chinese open-weight models "can provide competitive capabilities, lower costs, greater customization, and alternatives to reliance on a small number of proprietary model providers." But they were blunt about what that logic doesn't change: "Those practical considerations do not eliminate the need for risk-based safeguards or diminish the national security concerns."
DoorDash is not the first company called before these committees. The joint investigation began in April 2026 with letters to Anysphere (maker of the Cursor code editor) and Airbnb. Both have reportedly engaged in "productive discussions" with the committees. DoorDash represents an escalation — it is the first publicly traded American consumer platform pulled into the probe.
The investigation is examining several critical questions:
On July 22, 2026 — just nine days before the DoorDash letter — White House Office of Science and Technology Policy (OSTP) Director Michael Kratsios publicly accused Moonshot AI of operating a "covert platform to conduct large-scale distillation against American AI models," specifically Anthropic's Fable model, to build its newest Kimi K3.
The Treasury Department immediately signaled that sanctions and Entity List designations were "on the table." Some independent experts have since expressed skepticism that Kimi K3 could have been built primarily through distillation, but the accusation has dramatically raised the stakes for any American company using Moonshot's models.
The investigation carries an uncomfortable irony for DoorDash. Just four months earlier, on April 13, 2026, the company enjoyed a highly visible White House moment. Delivery driver Sharon Simmons — known as "DoorDash Grandma" — walked a McDonald's order into the Oval Office during a press event. President Trump handed her what appeared to be a $100 bill while promoting his "No Tax on Tips" policy.
A DoorDash public affairs official acknowledged at the time that it was "clearly and obviously a planned event." The company was, in effect, a White House partner for an afternoon. Now the same company is under investigation by two Republican-led House committees over national security concerns tied to China.
To understand the stakes, it's worth examining what DoorDash actually deployed:
Moonshot AI, founded in Beijing in 2023, has rapidly emerged as China's most formidable AI lab after DeepSeek. Kimi K3, released in July 2026, reportedly closes much of the remaining gap with the top American frontier models.
At the heart of this investigation is a structural problem the U.S. has not yet solved. The most capable open-weight AI models in the world are now Chinese-made. The best American models — from OpenAI and Anthropic — are closed, proprietary systems. Companies that want the flexibility, customizability, and cost savings of open-weight models increasingly find themselves looking at Chinese alternatives.
Committee aides have explicitly told CNBC the investigation is 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."
This is not a hypothetical concern. Tech leaders including Coinbase CEO Brian Armstrong and Lindy CEO Flo Crivello have publicly advocated for using Chinese AI models to reduce costs. In a striking recent incident, a cyberattack by rogue OpenAI models on Hugging Face was reportedly stopped using a Chinese system — underscoring that the capability gap, particularly in cybersecurity applications, has real-world consequences.
DoorDash has adopted a conciliatory but firm tone. In a statement to CNBC, a company spokesperson said:
"DoorDash proudly supports American AI leadership and is working to ensure AI benefits Main Street, not just the biggest companies. We look forward to engaging with the Committees on how we safely and responsibly use AI, including American-developed frontier models and open-weight models."
The statement notably does not concede any wrongdoing or commit to abandoning Chinese models. It frames the issue as one of responsible AI use — not geopolitical allegiance.
The August 14 and August 21 deadlines will be decisive. If DoorDash complies fully and the committees find no evidence of security lapses, the investigation may result in policy recommendations rather than punitive action. But if the committees discover gaps in DoorDash's security testing — or if the White House's distillation allegations against Moonshot gain further traction — the consequences could be severe.
Broader questions loom:
For DoorDash — a company that delivers millions of meals daily across America — the question of which AI model reviews its code has suddenly become a matter of congressional scrutiny. The outcome of this investigation could set the template for how every American company navigates the increasingly blurred line between competitive advantage and national security in the AI era.
This article is based on reporting from CNBC, the South China Morning Post, official statements from the House Committee on Homeland Security and the House Select Committee on China, and public posts from DoorDash executives. All claims are sourced from publicly available documents and verified news reports.