China Puts Robotaxi Expansion on Ice After Baidu’s Wuhan Meltdown

China just hit the brakes on its robotaxi ambitions — hard.

According to a Bloomberg report citing unnamed sources, the country has suspended issuing new licenses for autonomous vehicles. The immediate trigger? A mess in Wuhan last month where dozens of Baidu Apollo Go robotaxis simply stopped moving, creating a gridlock that made national headlines.

I’ve been watching the Chinese AV space for years, and this freeze feels less like a surprise and more like an overdue reality check. The Wuhan incident wasn’t just a minor hiccup; it was the kind of public failure that regulators in Beijing couldn’t ignore. When a fleet of driverless cars turns a major city intersection into a parking lot, the optics are terrible for an industry that’s been promising seamless autonomy.

So what’s actually frozen? Companies can’t add new driverless cars to their fleets, expand into new cities, or start new test projects. That’s a pretty comprehensive shutdown. The report doesn’t give a timeline for when officials might start issuing new licenses again, which tells me the review process could drag on.

I find it interesting that Beijing is specifically urging local governments to review the sector. That suggests the central authorities want more oversight — and probably stricter safety protocols — before anyone gets back on the road. The Wuhan incident exposed a gap in operational readiness, and regulators are now scrambling to close it.

This isn’t the end of robotaxis in China, but it’s a serious pause. Baidu has been the poster child for autonomous driving there, and this freeze puts a big question mark over their expansion plans. I’d expect competitors like Pony.ai and WeRide to feel the heat too, even if their tech wasn’t directly involved.

For now, the message from Beijing is clear: move fast and break things doesn’t fly when the things you break are traffic jams involving dozens of empty cars. The industry will survive, but this pause might actually be healthy — if it forces better testing and real-world fallback plans. We’ll see.