You order pad thai on a Tuesday night. A congressman calls it a national security threat.
That’s not a joke. That’s what just happened. US lawmakers have launched an investigation into DoorDash for using Moonshot AI’s Kimi K2.6 model — a Chinese-built AI — in its operations. The implication is clear: your food delivery app might be a Trojan horse for foreign surveillance.
Except it almost certainly isn’t. And if you look past the headlines, this probe has less to do with protecting your data and more to do with protecting someone else’s market share.
The scariest part of this story isn’t that a Chinese AI is inside an American app. It’s that the US government has decided it gets to pick which AI you’re allowed to use.
Here’s the tension nobody in Washington wants to talk about: American companies need the best AI tools to compete globally. Right now, some of the most cost-effective, high-performing models come from Chinese labs like Moonshot, DeepSeek, and others. They’re cheaper, they’re fast, and in several benchmarks, they’re competitive with or beating Western alternatives. That’s not propaganda — that’s just the market.
But the US government is in full decouple-from-China mode. Every Chinese tech company is treated as a potential arm of the CCP. Every data flow is a potential leak. And every American business that touches Chinese infrastructure is a potential traitor — or at least, a useful political target.
DoorDash isn’t shipping classified military data through Kimi K2.6. It’s optimizing delivery routes and maybe handling some customer service automation. The idea that your dumpling order is being funneled to Beijing is the kind of paranoid fantasy that plays great in a congressional hearing and makes zero sense in reality.
If the government can investigate your delivery app for using the wrong AI, there is no app, no tool, and no business decision that’s off-limits.
So what’s really going on? Follow the money — not to Beijing, but to Washington.
US AI firms are pouring record amounts into lobbying. OpenAI, Anthropic, Google DeepMind — they all have skin in this game. If American companies can freely choose cheaper, equally capable Chinese models, the pricing power of US AI providers collapses. Why pay premium subscription fees for a domestic model when a Chinese alternative does the job at a fraction of the cost?
This investigation sends a message to every CTO in America: use a Chinese model, and you might end up in a congressional probe. Use a US model, and you’re safe. That’s not security policy. That’s a competitive subsidy dressed up in a flag.
National security is the most powerful phrase in American politics because it’s the one argument that requires no evidence and permits no debate.
And that’s what makes this dangerous. There ARE legitimate concerns about data flowing through Chinese infrastructure. There ARE real questions about AI models trained on data that could include sensitive information. But those concerns deserve a serious, transparent regulatory framework — not a performative investigation into a food delivery company that happens to use a model lawmakers have never heard of.
If you use DoorDash, this affects you — not because your data is at risk, but because the tools that make your experience cheaper and faster are being politically curated. If you invest in AI, this matters because the competitive landscape is being shaped by congressional hearings, not market performance. If you work in tech, this matters because your architecture decisions are now geopolitical decisions.
The real question isn’t whether DoorDash should use Kimi K2.6. The real question is: who decides which AI you’re allowed to use — you, or your congressman?
When the government chooses your tools for you, security isn’t the goal. Control is.
FAQ
Q: Is there actually a real security risk with Chinese AI models?
A: There can be, depending on what data flows through them. But the DoorDash probe targets a food delivery app optimizing routes and customer service — not a defense contractor handling classified intel. The risk profile is wildly different, and lawmakers know that.
Q: What does this mean for companies choosing AI providers?
A: It means your AI stack is now a compliance decision, not just a technical one. If you pick a Chinese model, you risk regulatory scrutiny. If you pick a US model, you may pay more for equivalent performance. CTOs now need a geopolitical strategy alongside their tech strategy.
Q: Isn't this just protectionism disguised as security?
A: Largely, yes. The US AI industry benefits enormously from creating fear around Chinese alternatives. When lawmakers investigate a delivery app for using a Chinese model, it chills adoption across the entire market — effectively steering business toward domestic providers without an explicit mandate. It's a subsidy by intimidation.