Waymo is now providing over 400,000 rides per week across 10 US cities, while Elon Musk just pushed unsupervised Tesla FSD to Q4 2026 at the earliest. Meanwhile, Baidu's Apollo Go hit 300,000 weekly rides before a system malfunction in Wuhan triggered a nationwide permit freeze in China. After a decade of promises, self-driving cars in 2026 are finally reaching commercial scale, but the gap between leaders and laggards has never been wider.
| Company | Status | Weekly Rides | Cities Operating | Key Development |
|---|---|---|---|---|
| Waymo | Fully driverless | 400,000+ | 10 US cities | Raised $16B at $126B valuation, targeting 1M rides/week by year-end |
| Apollo Go (Baidu) | Fully driverless | 300,000 (peak) | Multiple China cities, Dubai, Abu Dhabi | Fleet halted in Wuhan after malfunction; China froze new permits |
| Tesla FSD | Supervised only | N/A | 0 (commercial) | Unsupervised pushed to Q4 2026; HW3 cars can't achieve full autonomy |
| Cruise | Limited operations | Minimal | Paused expansion | Still recovering from 2023 San Francisco incident |
| Zoox | Testing phase | Free rides only | San Francisco, Las Vegas | Targeting waitlist removal by year-end 2026 |
In February 2026, Waymo opened its fully autonomous ride-hailing service to the public in four new cities: Dallas, Houston, San Antonio, and Orlando. That brings the total to 10 US cities where you can hail a driverless Jaguar I-Pace from your phone. Waymo co-CEO Tekedra Mawakana stated the company is "on track to serve over one million rides per week by the end of this year", up from roughly 250,000 weekly rides in early 2026.
The expansion is backed by serious capital. Waymo closed a $16 billion funding round at a $126 billion valuation, the largest investment ever in an autonomous vehicle company. The round included Dragoneer Investment Group, DST Global, Sequoia Capital, and Andreessen Horowitz. Waymo has previously announced 2026 launch plans in six other US cities including Denver, Detroit, Las Vegas, Nashville, San Diego, and Washington DC, plus London.
The company is also tackling its biggest technical challenge: winter weather. Minneapolis, Denver, and Detroit signal Waymo believes it's nearly ready to crack autonomous driving's toughest nut: winter conditions, where camera sensors struggle and roadway markings disappear under slush. Waymo will expand its Arizona plant with manufacturing partner Magna to build more than 2,000 vehicles by the end of 2026, with plans to eventually produce tens of thousands annually using Zeekr's RT platform.
Waymo's data from over 127 million miles shows the company achieved a ten-fold reduction in serious injury or worse crashes and 12-fold reduction in injury crashes with pedestrians compared to human drivers. Independent analysis backs this up. Waymo's accident rate is 2.1 incidents per million miles driven, while human drivers have a rate of 4.85 incidents per million miles, making Waymo's rate 2.3 times lower.
But context matters. Of 2,052 AV incidents reported to the NHTSA, autonomous vehicles were solely at fault for only 4% of accidents that involved other road users. The vast majority of AV-caused injuries involve animals; only 14 human injuries resulted from the 234 incidents where an AV was at fault, with zero human fatalities caused by an AV from 2021-2025. According to the NHTSA through February 2026, fully self-driving car deaths were equal to only 2, while cars with ADAS2 (not fully automated) have been involved in 54 deaths.
The safety record isn't flawless. The NHTSA and NTSB recently opened investigations into incidents where Waymo robotaxis reportedly failed to stop for school buses in Austin and Atlanta, and an investigation was launched after a Waymo vehicle struck a child near a school in California, resulting in minor injuries. Still, these represent edge cases in a fleet that has now driven over 200 million fully autonomous miles.
Tesla's 2026 story is one of incremental software improvements and repeatedly pushed timelines. The company released FSD v14.3 in late April 2026, featuring improvements like a rewritten AI compiler resulting in 20% faster reaction time and unified models between Actually Smart Summon, FSD, and Robotaxi. But these remain supervised systems requiring constant driver attention.
During Tesla's Q1 2026 earnings call, CEO Elon Musk confirmed that unsupervised Full Self-Driving for consumer vehicles won't arrive until Q4 2026 at the earliest, saying "I'm just guessing here, but probably in the fourth quarter". Musk had promised full autonomy by 2018, a million robotaxis by 2020, and unsupervised FSD by June 2025, then moved the goalpost again in January 2026 when he admitted Tesla needed 10 billion miles of data before safely deploying unsupervised driving.
Worse news for early adopters: Musk confirmed that Hardware 3 vehicles "simply do not have the capability to achieve unsupervised FSD," citing that HW3 has "only one eighth of the memory bandwidth of hardware 4". For customers who paid thousands for FSD on HW3 vehicles, Tesla is offering discounted trade-ins to HW4 cars and a hardware upgrade path requiring replacement of both the computer and all cameras, which Musk acknowledged would require setting up "micro factories" in major cities.
Tesla is still only operating driverless Model Ys in Austin, and even that is at a tiny scale compared to Waymo's operations. The gap between Tesla's supervised driver assistance and Waymo's actual driverless service has never been wider.
Baidu's Apollo Go was on a tear before hitting a wall in April 2026. By February 2026, Apollo Go delivered over 20 million rides worldwide cumulatively, with 3.4 million fully autonomous trips in Q4 2025 alone, marking a year-on-year increase exceeding 200%. In its peak week, the company provided 300,000 trips, putting it nearly on par with Waymo.
Apollo Go was expanding aggressively internationally. The company partnered with Uber for commercial deployment in Dubai starting in February 2026, and through its partnership with Lyft, launched in Germany and the UK during 2026. Wuhan city is home to Apollo Go's largest robotaxi deployment in China, with more than 1,000 vehicles operating without human drivers.
Then came the Wuhan incident. On March 31, 2026, Apollo Go robotaxis stopped mid-traffic in Wuhan, leading to highway collisions, with preliminary findings from Wuhan traffic police suggesting system malfunctions as the cause. The fallout was swift. Chinese regulators stopped issuing new autonomous vehicle permits, preventing companies from adding additional vehicles, extending services to new areas, or launching new testing initiatives. Apollo Go's Wuhan robotaxi service remains shut down while local officials investigate, with no public timeline for when service might resume.
The story of self-driving cars in 2026 isn't about technology breakthroughs. The sensors work. The AI can navigate complex intersections. What's changing is operational scale, and that separates the companies that can actually run a robotaxi service from those still promising one.
Waymo operates over 2,500 vehicles providing hundreds of thousands of weekly rides. That requires fleet management infrastructure, remote assistance teams (Waymo employs about 70 remote assistance workers who can provide guidance when vehicles encounter edge cases), charging networks, cleaning operations, insurance frameworks, and regulatory relationships in a dozen jurisdictions. Building this operational machinery takes years and billions of dollars, which is exactly what Waymo's $16 billion raise enables.
Tesla, by contrast, has been promising a robotaxi network since 2019 but still doesn't operate a single commercial driverless route. The technical capability gap is real (camera-only systems vs. lidar-equipped vehicles), but the operational gap is wider. Running supervised FSD in customer-owned cars is fundamentally different from operating a fleet of driverless vehicles as a service business.
Apollo Go proved it could match Waymo's scale, but the Wuhan incident exposed the fragility of operating in a regulatory environment where a single system failure can halt an entire national industry overnight. The China permit freeze is a reminder that technical readiness doesn't guarantee commercial viability when governments can pull the plug at any moment.
The competition moving forward isn't about who has the best autonomous driving AI. It's about who can deploy capital efficiently, navigate patchwork regulations across jurisdictions, maintain safety records under regulatory scrutiny, and operate logistics networks at scale. On those measures, Waymo is years ahead, and the gap is widening.
For fully driverless systems like Waymo, yes. Waymo's accident rate of 2.1 incidents per million miles is 2.3 times lower than the human driver rate of 4.85 per million miles. Waymo also shows a 10-fold reduction in serious injury crashes compared to human drivers. However, this data comes from a limited operating environment (mostly fair-weather cities). The safety of supervised systems like Tesla FSD is harder to assess because the human driver is the ultimate safety backstop.
Waymo operates fully driverless ride-hailing in 10 US cities: Phoenix, San Francisco Bay Area, Los Angeles, Atlanta, Austin, Dallas, Houston, San Antonio, Orlando, and Miami. The company has announced plans to expand to at least 10 more US cities plus London by the end of 2026, though regulatory approval timelines vary by jurisdiction.
Elon Musk said "probably in the fourth quarter" of 2026 during Tesla's Q1 earnings call, but added this would be a gradual, geography-limited rollout. Tesla has repeatedly missed self-driving timelines (Musk promised a million robotaxis by 2020 and unsupervised FSD by June 2025). Tesla currently operates only supervised FSD, which requires constant driver attention, plus a tiny pilot of driverless Model Ys limited to Austin.
On March 31, 2026, multiple Apollo Go robotaxis stopped mid-traffic in Wuhan due to what police called system malfunctions, causing highway collisions. Chinese regulators responded by freezing all new autonomous vehicle permits nationwide, preventing companies from adding vehicles or launching in new cities. Apollo Go's Wuhan service remains suspended pending investigation, with no timeline for resumption.
No. Elon Musk confirmed in April 2026 that Hardware 3 vehicles cannot achieve unsupervised Full Self-Driving because HW3 has only one-eighth the memory bandwidth of Hardware 4. Tesla is offering discounted trade-ins to HW4 vehicles or a hardware upgrade path that requires replacing the computer and all cameras. This affects customers who paid up to $15,000 for FSD on HW3 vehicles based on promises of future full autonomy.
Waymo pricing is comparable to Uber and Lyft in most markets where it operates. In cities like Phoenix and San Francisco, a typical Waymo ride costs roughly the same as an UberX. Waymo also operates through the Uber app in Atlanta and Austin, where you can select Waymo as a ride option. Zoox currently offers free rides as it builds its network, but plans to introduce paid service.
For fully driverless systems like Waymo, yes. Waymo's accident rate of 2.1 incidents per million miles is 2.3 times lower than the human driver rate of 4.85 per million miles. However, this data comes from limited operating environments.
Waymo operates fully driverless ride-hailing in 10 US cities: Phoenix, San Francisco Bay Area, Los Angeles, Atlanta, Austin, Dallas, Houston, San Antonio, Orlando, and Miami. The company plans to expand to at least 10 more cities plus London by year-end.
Elon Musk said "probably in the fourth quarter" of 2026, but this would be gradual and geography-limited. Tesla has repeatedly missed self-driving timelines and currently operates only supervised FSD requiring constant driver attention.
On March 31, 2026, multiple Apollo Go robotaxis stopped mid-traffic in Wuhan due to system malfunctions, causing collisions. Chinese regulators responded by freezing all new autonomous vehicle permits nationwide, suspending Apollo Go's Wuhan service indefinitely.
No. Musk confirmed in April 2026 that HW3 vehicles cannot achieve unsupervised FSD because they have only one-eighth the memory bandwidth of Hardware 4. Tesla is offering discounted trade-ins or hardware upgrades requiring computer and camera replacement.
Waymo pricing is comparable to Uber and Lyft in most markets. A typical Waymo ride costs roughly the same as an UberX. Waymo also operates through the Uber app in Atlanta and Austin.