This sector has a genuinely different anchor structure from a chip-allocation-style industry with one private gatekeeper. There is no single company that decides who is allowed to sell LiDAR the way Nvidia decides who is allowed to build with its chips. Instead, demand is anchored by two more diffuse forces — individual automaker design-win decisions, and two national regulators moving at very different speeds and with very different specificity — and the honest finding of this section is that those two regulatory anchors currently pull in different directions.
Every automotive design win is its own multi-year negotiation: an OEM's engineering team tests, selects and qualifies a specific supplier's part against its own platform's sensor-fusion architecture, camera and radar suite, and in-house software stack (§1, §3). There is no industry-wide "LiDAR allocation" decision the way there is a chip-allocation decision in AI compute — but the same qualification-calendar moat applies at the level of each individual OEM relationship, and it is why a design win, once secured, tends to last the life of a vehicle platform (typically 5-7 years) regardless of whether a rival's part becomes technically superior in the interim.
China's GB 44721-2026 names the hardware. MIIT's certification-standards work for L3-level automated driving explicitly includes LiDAR among the hardware configurations a certified system must carry, alongside high-precision maps and compute — a direct, sourced link between a national regulatory instrument and demand for this specific sensor class. Combined with a genuinely fast-scaling robotaxi commercial rollout (Apollo Go's 100 million-plus cumulative rides, §6), China is currently the clearest, most concrete anchor-buyer story in this industry — and the design wins it is generating (RoboSense's exclusive Apollo Go contract, §6) are, as of this report's research date, going disproportionately to non-US-listed suppliers.
The United States' FMVSS 127 names an outcome, not a sensor. NHTSA's rule requires pedestrian automatic-emergency-braking performance up to 62mph by 1 September 2029, and that performance bar is genuinely difficult to meet with camera-and-radar alone at higher speeds — which is why several OEMs are reported to be adding LiDAR or upgrading radar specifications to comply — but the rule itself is sensor-agnostic, and is currently the subject of a proposed rulemaking (23 March 2026 DOT status filing) that would extend the compliance lead time by two years. Treat this as real but soft demand pull, not a confirmed hardware mandate, and watch the rulemaking's outcome as one of this report's numbers to track (§10).
Taken together, the two regulatory anchors are not additive, and a reader who assumes "regulation is pushing LiDAR adoption globally" is collapsing two very different stories into one. China's is a named, dated, hardware-specific requirement tied to a fast-scaling commercial rollout; the United States' is a sensor-agnostic performance rule whose own compliance timeline is currently in flux. This divergence is itself one of this report's more important findings, and it should shape how much regulatory tailwind a reader assumes behind any individual company's US-market growth story in the company reports that follow.
