China's 2030 Self-Driving Mandate: Own the Sensors, Not the Automakers

China's new five-year plan, published this week, puts a date and a denominator on the self-driving story: electric and "intelligent connected" vehicles should reach 70% of new passenger-car sales and 40% of new commercial vehicles by 2030, with vehicles "equipped with autonomous driving capabilities" deployed at scale. The trade that headline tempts is clean and wrong — buy the Chinese carmakers chasing the mandate. The counterintuitive part of the math is that it points the other way, to the component maker every level of autonomy depends on. And that stock has been falling while its business compounds.

The mandate is real this time

The policy is more than a slogan this time. Alongside the sales targets, the plan calls for highly automated driving on highways, urban expressways, and some city roads by 2030, plus commercialization of robotaxis, autonomous buses, and robotrucks. The regulatory plumbing is catching up to the ambition: Beijing issued its first Level 3 road-use permits in late 2025, proposed a dedicated autonomous-vehicle chapter in the national Road Traffic Safety Law in August, and enacted a mandatory Level 3/4 safety standard. That is a mandated ramp, not a promotional one. By one estimate, roughly 70% of new cars sold in China already carry Level 2 driver assistance, heading toward a projected 90% by 2030.

The wrong face of the story

The trap is picking the wrong face of the story. The Chinese automakers that dominate the headlines are in a brutal, government-funded price war, and the plan itself concedes it: it calls for phasing out weak, inefficient capacity and cracking down on unauthorized subsidies and price competition. The poster child of the trade, XPeng, is down about 49% this year and sits at its 52-week low — a falling knife, not a discount. "Buy the dip" on a consolidation-bound automaker is the reflex this company's own kill list exists to refuse.

Own the sensors, not the sticker

The leverage sits upstream. Every self-driving system, from Level 2 up, needs sensors to see, and the components scale with the mandate regardless of which automaker survives. That puts a LiDAR maker called Hesai in the path of the policy. Hesai is the global leader in automotive lidar with roughly a third of the market. The numbers justify the position: in the second quarter it shipped 628,275 units, up 78% from a year earlier, drove revenue up 22% to about $127 million, and posted net income of $10.4 million — its fifth straight profitable quarter. Robotics sensor shipments tripled. Full-year guidance points to 38% to 45% revenue growth in Q3.

Now the divergence that makes this worth writing about. Hesai trades at roughly 5.6 times trailing sales with a market cap of about $2.7 billion and a net cash balance sheet. Its unprofitable American rival Ouster trades at more than twice that sales multiple. Growth at a cheaper price than a money-losing peer is a GARP-with-teeth setup on paper.

The one honest flaw

So why is Hesai down about 22% this year and roughly 40% off its 52-week high? Not because the business slowed — the opposite. The discount is geopolitical, and it is the one honest flaw in the setup. blacklisted Hesai in 2024 as a Chinese military company, a designation the company lost a lawsuit over and is still appealing. Its sensors sit in American robotaxis, trucks, and airports, and it just expanded a partnership to be an option on Nvidia's autonomous-vehicle platform — which is precisely why U.S. lawmakers proposed legislation in May to phase out Chinese-made lidar. A hard U.S. phase-out would remove a meaningful piece of higher-value demand, and that risk is real, structural, and not fully priced.

That is the entire argument in one breath. The market has lumped Hesai into one "Chinese autonomy" risk bucket with the collapsing automakers, and priced it as if the whole franchise were a geopolitical bet. But the core it actually serves — the Chinese ADAS and autonomy ramp that the new plan mandates — is profitable, accelerating, and cheap. The overcorrection is on the China side; the genuine tail risk is on the U.S. side. Neither cancels the other.

That makes this a watch with a defined trigger, not a reflexive buy the dip. The math clears the bar on its own: a profitable, net-cash compounder growing units ~78% at roughly 5.6x sales, half the multiple of a loss-making peer, against a now-mandated demand base that was previously just aspirational. The unresolved variable is U.S. export policy, and it is a hard one. If the phase-out legislation dies or the discount keeps ignoring the China mandate, the stock re-rates toward the growth already sitting in the model. If it lands, the discount is earned. Either way, the edge is knowing the cheap multiple and the falling stock are two different things — one is the math, and the other is a risk still being litigated.