Industry paid a federal lab to test Chinese lidar. The buyers weighing it say they didn't know.

The funding arrangement is the load-bearing detail. A Department of Energy lab is testing parts, and the bill is being paid by a company or a group of companies in the electric and autonomous vehicle business, according to what the lab told two people familiar with the effort ^[3]. The result is a finding owned by whoever wrote the check, produced at an institution that told TechCrunch in June and again in July that this was not a project it was at liberty to discuss with reporters ^[6]. A published federal safety finding is a rule everyone can plan around. A privately commissioned one is leverage.

Then there is the awkward part. Rivian, General Motors, Ford, Kodiak, Lucid, Nuro and Uber are all reported to be considering Chinese lidar, and all told TechCrunch they had not heard of the review ^[4]. Nvidia, Zoox and Aurora did not answer the question ^[5]. That is ten firms in the frame with nobody claiming the work ^[20]. Either the funder sits outside that list, or someone inside it is buying due diligence quietly enough that peers evaluating the same sensors cannot use the answer. Omer Keilaf, who runs the Israeli lidar maker Innoviz, said he also did not know about the review until TechCrunch raised it ^[11].

Keilaf supplied the most detailed threat model in the reporting, and he sells the alternative. His scenarios are sensors disabled at scale, disrupting vehicles in motion and bricking the ones parked ^[13], and lidar's dense returns compressed and pushed out over a cellular chip, which he said his seven years in Israeli intelligence tell him is not difficult to do ^[14][15]. The first has only partial precedent: researchers have disabled individual sensors with lasers, and how a fleet-wide disruption of robotaxis in active use would work is not obvious ^[16]. He also named a risk requiring no covert intent, which is that Beijing has deemed lidar strategically important, so an autonomy stack built on Chinese supply is exposed on its own terms ^[12].

According to the same two people, the review may end up clearing the sensors on cybersecurity ^[10]. That outcome is the one that shows how little the technical question now decides. The bills from Baldwin and Budd and from Moolenaar are built on espionage and back-door arguments ^[8][9], and a statute does not read lab reports. The commercial pull runs the other way: lidar started far above radar and cameras on price and has fallen sharply over the past decade ^[17][18], which is why anyone is looking at Hesai and RoboSense in the first place, and neither company responded to TechCrunch ^[19].

None of this has to resolve to change a sourcing decision. A sensor that might be prohibited in two years ^[7], whose only serious American examination is unpublished and belongs to a party you cannot identify, does not need to fail the test to lose the socket. It needs one program manager to decide the part is not worth the paperwork.