Autonomous vehicle.
The robotaxi factory race.

Waymo's newest robotaxi is built in China. The company that proved self-driving works is shipping its cars across the Pacific and paying tariffs on them, because that's where the cheap cars are.

That tells you where the bottleneck has moved. The driving is solved well enough. Waymo now gives more than 500,000 paid rides a week, and its own data shows 82% fewer injury crashes than human drivers. Tesla runs robotaxis in seven metro areas, and Chinese operators are expanding abroad on Uber's network. The hard part now is building self-driving cars cheaply enough, and quickly enough, to fill a fleet.

That's the race autonomous vehicle stocks are running in 2027: cost per car, cars on the road, and keeping regulators onside while the fleets multiply.

The autonomous vehicle stocks below start with the leader, then the cost cutters, a cautionary tale, the ride-hail networks, the freight bet, and two private names.

One company has already answered the safety question at scale, which leaves it with a manufacturing problem.

Alphabet NASDAQ: GOOGL

Alphabet owns most of Waymo, which raised $16 billion at a $126 billion valuation in February. Waymo went from a quarter of a million weekly rides in April 2025 to more than 500,000 a year later. Its stated goal is a million a week by the end of 2026.

Getting there is mostly a matter of putting more cars on the road. That's why its new Zeekr-built vehicle matters more than any software update, and why Waymo opened Dallas and Houston to the public in August. Alphabet stock is a diluted way to own all this, but it's the only public way to own the leader.

These four companies are attacking the bill of materials, and whoever gets a driverless car cheapest gets to build the biggest fleet. For autonomous vehicle stocks, that's the race that matters most.

Tesla NASDAQ: TSLA

Tesla is betting that cameras and a giant factory beat lidar and a careful rollout. Robotaxi is live in seven metros, and it has started running without a safety monitor in parts of Texas and Florida. In the second quarter, Cybercab production started at Gigafactory Texas, which Tesla lists with more than 125,000 units of capacity.

Nobody else on this list can build cars at that scale. What Tesla doesn't have yet is Waymo's track record: its robotaxis have logged a few million paid miles, against Waymo's 270 million driverless ones. If the camera-only approach holds up as it scales, Tesla has the cheapest robotaxi on the market. If it doesn't, the factory doesn't matter.

Pony AI NASDAQ: PONY

Pony AI has the cleanest cost story in the group. Its seventh-generation self-driving kit costs 70% less to build than the last one. It's targeting a total vehicle cost under 230,000 yuan by 2027, about $34,000 for a car that drives itself.

Cheap cars mean more cars. Pony ended June with 1,975 robotaxis and is aiming for more than 3,500 by year-end, including a deal to put over 2,000 on Uber in Europe. Robotaxi revenue rose almost eightfold from a year earlier.

WeRide NASDAQ: WRD

WeRide treats regulation as a map. It runs fully driverless in Dubai and Abu Dhabi, operates in more than 60 cities across 13 countries, and is lining up Madrid and Zurich with Uber. Overseas revenue rose 164% last quarter.

Hesai NASDAQ: HSAI

Hesai shows how fast self-driving hardware is getting cheaper. Lidar used to be the part that made robotaxis absurdly expensive. Hesai shipped 1.1 million units in the first half of 2026, double a year earlier, and it was the first lidar maker to turn a full-year profit.

Scale cuts both ways, and this is what happens when the public stops trusting the fleet.

Baidu NASDAQ: BIDU

Baidu ran what may have been the world's busiest robotaxi service. Its Apollo Go fleet hit more than 350,000 driverless rides a week in March. Then, on March 31, more than 100 cars stalled across Wuhan and stranded passengers for up to two hours.

Baidu suspended service in the city. Its filings imply rides fell roughly 70% in the second quarter, and the latest release skipped the ride count entirely. Apollo Go still runs in 28 cities, including driverless service in Dubai, so this is now a turnaround: can a robotaxi leader win back a city after a night like that?

Robotaxi operators need riders, and the ride-hail apps already have them. The fight is over who owns the customer.

Uber NYSE: UBER

Uber is signing up every fleet it can. Partners have committed about 120,000 autonomous vehicles to its network, and it expects to be live in as many as 15 cities by the end of the year, up from seven.

The exception is the one that matters most. Waymo ended its Phoenix partnership with Uber in July and plans to launch its own app in Austin and Atlanta in 2028. Uber is becoming the storefront for everyone except the leader.

Lyft NASDAQ: LYFT

Lyft found a less glamorous way in: running the fleets. Since June it has been handling fleet operations for Waymo in Nashville, the cleaning, charging and maintenance a car with no driver still needs. A new 80,000-square-foot depot is due to open this month.

Every robotaxi needs somewhere to sleep. If operators would rather rent that job out than build it, Lyft has a business that doesn't depend on owning the app.

Trucking skips the hardest part of robotaxis, which is crowded city streets, and goes after a driver shortage instead.

Aurora Innovation NASDAQ: AUR

Aurora runs driverless trucks on highways across Texas and the Southwest and has logged about 440,000 driverless miles with no collisions it was at fault for. Its second-generation hardware costs half as much, and it plans to end the year with 200 driverless trucks.

The catch is that Aurora barely has revenue yet. It guides to $14–16 million of revenue this year against roughly $225 million of cash burn a quarter. Aurora is a bet that trucking economics show up before the cash runs out.

Neither trades on its own yet. One has a regulatory first, and the other has a listing on the way.

Zoox

Zoox, owned by Amazon, builds a pod with no steering wheel or pedals. In July it got the first federal exemption that lets a vehicle like that carry paying customers, for up to 2,500 vehicles a year. It started charging fares in Las Vegas in August.

May Mobility

May Mobility is going public through a merger with a blank-check company, at a value of about $1.4 billion, and will trade as MAY. It has given more than 550,000 commercial rides and partners with Uber, Lyft and Grab. It's small, unproven at scale and about to become the cheapest public way to buy a pure robotaxi operator.

The cheapest robotaxi hardware in the world is Chinese. Pony, WeRide and Baidu build on Chinese supply chains, and even Waymo buys its newest cars there.

The U.S. has already started closing that door. Federal rules bar Chinese driving software from connected cars sold here starting with 2027 models, and Chinese connected-car hardware from 2030.

So the robotaxi market, and the autonomous vehicle stocks riding it, is splitting in two. In the U.S., fleets will be built from domestic or allied parts at a higher cost per car, by Waymo, Tesla and Zoox. Everywhere else, from the Gulf to Europe to Southeast Asia, Chinese operators will show up with cars that cost a fraction as much, usually on Uber's app.

That means American robotaxis may end up the expensive ones. The U.S. invented the technology, but the cheapest fleets in the world will run somewhere else.