For years, self-driving cars were a promise stuck in permanent beta. Every January brought a new round of predictions, and every December came and went without the fully autonomous future anyone was sold. That cycle has finally broken. Robotaxis are now carrying paying passengers across multiple continents, several million miles a week, with real money changing hands and real companies fighting over market share. The technology isn't finished, and it certainly isn't perfect, but it has crossed the line from demo to business. What's worth understanding now isn't whether robotaxis are coming. It's who's actually winning, how safe the cars really are, and what the next eighteen months look like.
Waymo Built a Lead That's Hard to Ignore
Waymo operates roughly 3,000 robotaxis providing close to 500,000 paid rides a week and around 4 million rider-only miles weekly across more than a dozen US cities. That scale didn't happen overnight. Waymo spent years mapping cities block by block, running vehicles with safety drivers, then with employees as passengers, before opening up to the public. It's slow and expensive, but it built trust with regulators and city officials along the way, which is now paying off.
Waymo is targeting more than a million rides a week by the end of 2026, a level it has already touched on a monthly basis since spring. The company is also reportedly raising money at a valuation near $100 billion, roughly double where it stood a year earlier. Waymo is permitted or testing in 26 of the top 30 US metro areas and has plans to launch in 12 more cities this year, including London. A chunk of that expansion is happening through partnerships rather than Waymo going it alone. Uber is helping manage cleaning, charging, and maintenance in cities like Austin and Atlanta, while Lyft is set to support a Nashville launch and Avis Budget Group is lined up for Dallas. Spreading operational work across partners lets Waymo focus on the driving software while someone else worries about fleet logistics.
It isn't all smooth sailing. Sundar Pichai has reportedly told people internally that the robotaxi business won't be financially meaningful to Alphabet until 2027 or 2028, which is a useful reminder that even the market leader is still running this at a loss.
Tesla's Bet Looks Very Different, and It's Behind Schedule
Tesla took the opposite approach. Instead of painstakingly mapping cities, it bet that cameras, neural networks, and a much lighter prep process could get robotaxis to scale faster and cheaper. So far, the results haven't matched the ambition. Elon Musk projected 500 robotaxis in Austin by the end of 2025. That target wasn't met, and by mid-2026 Tesla had expanded to seven metros with only around 44 active vehicles total, a far cry from the nationwide fleet Musk has talked about for years.
Reporters who've actually tried to hail a Tesla robotaxi in newer markets like Houston and Dallas have described long waits and a fleet so small it barely counts as a service yet. Tesla's wait times run around 15 minutes or more, compared to roughly 5.7 minutes for Waymo. Pricing tells a different story though. Tesla charges around 81 cents a mile against Waymo's $1.36 to $1.43, which suggests Tesla is betting heavily on cost as its long-term edge once the fleet actually grows.
There's also a structural wrinkle. Most Tesla robotaxis on the road today still carry a human safety monitor, which means the company hasn't actually proven unsupervised operation at any real scale yet. Its core Full Self-Driving software still runs under driver-assistance rules rather than a true autonomous vehicle framework, so a lot of Tesla's path forward depends on regulatory changes that haven't fully arrived.
It's Not Just an American Story
While Waymo and Tesla get most of the headlines in the US, the robotaxi race is genuinely global. Baidu's Apollo Go is already running across more than ten Chinese cities, including Wuhan, Beijing, Shanghai, Shenzhen, Chongqing, and Guangzhou, with ride volumes in its biggest markets comparable to Waymo's on a per-city basis. Pony.ai and WeRide are scaling alongside it. Amazon's Zoox launched commercial service in late 2025 with its own purpose-built, steering-wheel-free vehicle. Mobileye is preparing to launch its own autonomous ride-hailing service after years spent supplying driver-assistance tech to traditional automakers, and Uber keeps stacking AV partnerships on top of its existing ride-hailing network, including a Nuro-powered Lucid Gravity fleet headed for Houston.
The takeaway is that no single company owns this category yet. China is moving at serious scale with its own players, the US is split between Waymo's methodical rollout and Tesla's aggressive but bumpy one, and a handful of smaller companies are carving out niches rather than trying to out-build Waymo directly.
The Safety Numbers Are More Nuanced Than the Headlines
Robotaxi crashes get a lot of attention, and the raw numbers can look alarming out of context. Tesla's Austin fleet logged 14 crashes in about 800,000 miles, roughly one every 57,000 miles, which on paper looks worse than typical human driving rates. But the detail matters. A significant share of those incidents were rear-endings and sideswipes caused by inattentive human drivers hitting a stopped or slow-moving robotaxi, not the other way around, and Waymo reports a similar pattern in the majority of its own incidents. Of 17 total reported Tesla incidents through March 2026, there were zero major crashes, and only six involved the autonomous system being at fault or partially at fault. Two of the more serious incidents actually happened when a remote teleoperator took manual control and made a mistake, not when the software itself was driving, which says something interesting: the handoff between machine and human is its own risk, separate from how good the underlying driving software is.
Waymo's safety record has held up reasonably well at much larger scale, though it isn't crash-free either. The honest read here is that robotaxis aren't yet dramatically safer than good human drivers across the board, but they also aren't the disaster some headlines suggest. They're a new category with a new failure pattern, and it's going to take more miles and more transparency from every company before anyone can make a confident comparison.
Regulation Is Catching Up, Unevenly
Every state and country is writing its own rules, which is creating a patchwork that companies have to navigate city by city. Texas passed SB 2807, a statewide authorization framework for commercial autonomous vehicles that becomes fully enforceable on May 28, 2026, covering safety and insurance compliance requirements. Other markets, like Washington DC, are still working through whether the right regulations will even be in place in time for planned launches. London is set to get its first Waymo deployment this year too, which will be a real test of how a tightly regulated European city handles American-style robotaxi expansion.
What This Actually Means Going Forward
The honest summary is that robotaxis have moved from science experiment to actual industry, but the gap between the leader and everyone else is still wide. Waymo has the scale, the safety record, and the regulatory relationships. Tesla has the cost structure and the ambition, but needs years of execution to catch up on fleet size and proven safety. China is quietly running its own massive operation that most Western coverage barely mentions. None of this is settling down anytime soon, and the next year of expansion, crash data, and city-by-city regulation will tell us a lot more than any keynote presentation has.
Keep following AutoVoB for more insightful automotive content.
Sign in to leave a comment.