China restarted robotaxi permits and its companies are racing into Dubai, Abu Dhabi, and Singapore. Cost matters: Baidu builds vehicles for $29,000 using EV supply chains while US firms refine premium markets at home. Strategy follows economics.
Good Morning from San Francisco, Microsoft had an interesting week on the transparency front. Australia's competition watchdog sued,
Australia's regulator says Microsoft pushed 2.7 million subscribers toward pricier AI plans by hiding a cheaper option inside the cancellation flow. The case tests whether essential software providers must disclose alternatives as clearly as upgrades.
China restarted robotaxi permits and its companies are racing into Dubai, Abu Dhabi, and Singapore. Cost matters: Baidu builds vehicles for $29,000 using EV supply chains while US firms refine premium markets at home. Strategy follows economics.
China has quietly restarted robotaxi permits after a months-long pause, and its champions—Baidu, WeRide, and Pony.ai—are moving fast into Dubai, Abu Dhabi, and Singapore while Waymo focuses on a handful of U.S. cities.
The shift follows a freeze amid taxi-driver protests in late 2024; permits are now flowing again, a decision consistent with Beijing’s plan to lead autonomous vehicles by 2035, according to a recent robotaxi permit resumption analysis. It’s a calculated bet.
Key Takeaways
• China lifted a six-month robotaxi permit freeze, prioritizing 2035 autonomous vehicle leadership over taxi driver job concerns
• Baidu's RT6 costs $28,800 using EV supply chains—Chinese firms target volume markets while US companies focus on premium cities
• Baidu, WeRide, and Pony.ai signed Middle East deals after GM's Cruise shut down Dubai operations in late 2024
• WeRide expects per-vehicle profitability at fleet scale in Middle East, testing if cost advantages create sustainable business models
Baidu’s Apollo Go now runs driverless in Wuhan. WeRide operates without safety drivers in parts of Guangzhou. Pony.ai is taking passengers in Shenzhen and building toward a four-digit fleet. All three signed Middle East deals within weeks of GM-backed Cruise shutting down its Dubai program late last year. The door opened—and they walked through it.
The regulatory turn matters more than the headline suggests. China paused approvals for several months in the second half of 2024 when labor pushed back. That pause ended quietly in 2025. The signal: strategic competition outweighs short-term job concerns. That choice is telling.
Costs amplify the impact. Baidu’s RT6 robotaxi is built for roughly $29,000 using China’s EV supply chain and locally sourced sensors. Waymo doesn’t publish unit costs, but analysts peg its hardware far higher; Tesla’s promised “sub-$30,000” Cybercab is still not in market. The upshot is not just cheaper cars—it’s a different go-to-market. Chinese operators can aim for break-even with hundreds of vehicles in value-oriented cities. U.S. rivals need premium routes and near-perfect execution to justify their bill of materials. Geography becomes strategy.
Dubai wants a quarter of trips autonomous by 2030. Abu Dhabi targets a similar share by 2040. Saudi Arabia’s Vision 2030 calls for a national autonomous network. The climate helps. Reliable weather, new roads, and supportive regulators make the Gulf a clean proving ground.
Labor politics also differ. In Gulf cities, much of the taxi workforce consists of guest workers with limited political clout. Pushback is less intense than in China or the U.S., lowering deployment friction. From Dubai’s view: Chinese robotaxis arrive turnkey and cheap. From Beijing’s lens: overseas rollouts validate the tech and build diplomatic capital. From Waymo’s perspective: defending core U.S. markets matters more than chasing small, early-stage volumes abroad.
Partnerships are locking in. Uber is working with WeRide in Abu Dhabi. Lyft has tied up with Baidu for potential European launches, pending approvals. In Singapore, Pony.ai and WeRide plan limited services with Grab and ComfortDelGro. The Chinese firms aren’t waiting for American entrants; they’re planting flags while U.S. players refine service in Phoenix, San Francisco, and Los Angeles. Speed compounds advantage.
China is replaying its EV playbook: long-horizon capital, vertical integration, and a strategic-sector mandate. U.S. development remains market-driven—fundraising cycles, public scrutiny, and fewer policy cushions when technology stumbles. Scale begets scale.
On talent and tech, both sides can drive in geofenced urban cores; neither claims victory in heavy snow or messy edge cases. The difference shows up in cost and iteration. When an RT6 can carry nearly 40 sensors—eight lidar units and a dozen cameras—at sub-$30,000, operators can flood zones, collect data faster, and tune software on real streets. Volume is a feature, not a byproduct.
Industry academics note that “who’s ahead” is situational. China leans on HD maps, roadside sensors, and 5G links in pilot zones; U.S. programs emphasize vehicle self-reliance. But while philosophies diverge, China’s manufacturing base compresses hardware prices and shortens upgrade cycles. That’s a structural edge in a field where hardware is still half the battle.
The business question is blunt: can robotaxis clear operating costs without permanent subsidies? WeRide’s CFO says per-vehicle profitability becomes realistic at fleet scale in the Middle East. Now comes proof.
Watch three markers. First, whether Dubai or Abu Dhabi hits break-even routes by 2026 with Chinese fleets. Second, whether leading operators push beyond 1,000 vehicles each without reliability regressions. Third, which European regulators greenlight limited service—Germany, the U.K., or a smaller market first. Those signals will show whether low cost plus policy support translates into durable unit economics.
U.S. players face a different bind. Until their component costs fall, global expansion remains a high-burn proposition. Chinese firms, riding domestic EV capacity and cheaper sensors, can treat the Gulf as a scale lab and Europe as the credibility prize. If they prove out per-car margins abroad, they return home with leverage—on regulators, on partners, and on pricing.
China’s EV ascent took decades: batteries, motors, cameras, and lidar brought in-house, then scaled. Autonomous vehicles are drafting behind that build-out. The U.S. led with software polish. China matched competence and undercut price. That’s why Chinese robotaxis are showing up where American brands aren’t—and why they may get to sustainable economics first.
Q: Why did GM shut down its Cruise robotaxi program in Dubai?
A: GM closed Cruise entirely in late 2024 after safety incidents damaged confidence and costs mounted. Dubai had an exclusive partnership with Cruise starting in 2023, but when GM pulled the plug on the whole business, the emirate quickly signed deals with Baidu, WeRide, and Pony.ai to fill the gap.
Q: How do Chinese and US autonomous driving approaches differ technically?
A: Chinese systems rely heavily on HD maps, roadside sensors, and 5G connections in pilot zones—infrastructure helps the car navigate. US programs like Waymo emphasize vehicle self-reliance, where onboard sensors and AI handle everything. Neither approach handles heavy snow or complex edge cases well yet.
Q: Why is the Middle East better for testing robotaxis than other regions?
A: Consistent sunny weather eliminates sensor problems from rain or snow. New road infrastructure has clear lane markings and predictable layouts. Most importantly, taxi drivers are guest workers with limited political influence, so labor opposition stays minimal compared to China or the US.
Q: When will Chinese robotaxis launch in Europe?
A: Lyft and Baidu announced plans to launch robotaxi services in Europe starting in 2026, pending regulatory approval. No specific cities are confirmed yet. Germany and the UK are likely targets, but smaller markets with friendlier regulations might approve services first as proof-of-concept launches.
Q: How many robotaxis does Waymo operate compared to Chinese companies?
A: Waymo operates fully driverless cars in five US cities but doesn't publish fleet size. Pony.ai is building toward 1,000 vehicles by year-end. Combined, Baidu, WeRide, and Pony.ai have more robotaxi projects at commercial stages than US companies, though exact vehicle counts aren't disclosed.
Microsoft buried OpenAI losses in a $4.7B catch-all line without disclosing stake size, fair value, or revenue mechanics—despite a partner now worth $500B. Investors can't map exposure or returns. The gaps are fixable and overdue.
OpenAI's company knowledge mode connects workplace apps to ChatGPT—but the real test is whether enterprises will expose their entire institutional memory to AI. The feature points toward governed knowledge bases, yet arrives with manual toggles and gaps Microsoft solved months ago.
Anthropic secured Google's largest chip deal—up to 1M TPUs worth tens of billions—while keeping Amazon as primary partner. The rare multi-cloud strategy gives the startup leverage both clouds typically demand for themselves. Can neutrality scale?
OpenAI bought the team behind Sky, a Mac assistant that actually controls your apps. It's a move from answering questions to doing work—and puts pressure on Apple's automation story. Integration timing and privacy details will decide if it sticks.
Get the 5-minute Silicon Valley AI briefing, every weekday morning — free.
