Signals Inbox·July 21, 2026·Autonomous Systems

Waymo vs Zoox: which robotaxi model will win?

Waymo is winning the robotaxi race today and has the model most likely to win overall. Zoox built the more elegant vehicle, but Waymo has already built the network around it.

Q1Why are Waymo and Zoox compared so often?

Waymo and Zoox are compared because they want to dominate the same robotaxi market with opposite operating models. Waymo is scaling an autonomous driving system through modified production vehicles. Zoox is building the vehicle, software and passenger experience as one closed system.

Waymo’s approach is pragmatic. It removes the driver from cars that regulators, manufacturers and passengers already understand. Its current fleet is largely based on the Jaguar I-Pace, while its next generation will use vehicles such as the Hyundai Ioniq 5 and the purpose-developed, Zeekr-built Ojai platform.

Zoox started from the opposite question: what would a taxi look like if it had never needed a human driver? Its vehicle has no steering wheel or pedals, travels in either direction and seats four passengers facing one another.

The rivalry became commercially meaningful when Zoox launched public rides in Las Vegas and began opening access in San Francisco. Until then, Waymo had a robotaxi network while Zoox had an ambitious vehicle. Now they finally compete for passengers, cities, manufacturing capacity and regulatory permission.

Q2Why is it difficult to say whether Waymo or Zoox has the better model?

Waymo is clearly winning the robotaxi business today, but Zoox may still have designed the more efficient robotaxi for a mature market. The conflict is between demonstrated scale and unproven architectural upside.

Waymo has the evidence that matters now: paid rides, thousands of vehicles, multiple cities, repeated passenger usage and hundreds of millions of driverless miles.

Zoox’s case is mostly forward-looking. Its bidirectional vehicle could waste less time turning around. Its cabin uses almost the full footprint for passengers. Its 133 kWh battery is designed for more than 16 hours of operation.

Zoox has not yet shown that these features produce lower costs, higher utilization or faster city launches. A technically cleaner vehicle does not automatically create a stronger transportation network.

We therefore need two scoreboards. Waymo leads on what has already been proven. Zoox must show that its custom architecture changes the economics enough to compensate for entering the market years later.

Q3How much larger is Waymo than Zoox today?

Waymo is so far ahead of Zoox in commercial robotaxi scale that the two companies barely sit in the same category.

Alphabet said during its first-quarter 2026 earnings call that Waymo had exceeded 500,000 paid autonomous rides per week. That represents roughly 26 million annualized rides at the current rate.

Waymo’s co-CEO Dmitri Dolgov also referred to a fleet of approximately 3,000 vehicles. Dividing weekly rides by fleet size gives an average of about 167 rides per vehicle per week, or almost 24 rides per day.

Zoox has not disclosed a comparable paid-ride figure. Its services remain free or restricted in its first markets. A federal recall filed in July 2026 covered 105 purpose-built Zoox vehicles, providing one of the best public indications of its current fleet size.

Waymo therefore operates roughly 29 times more vehicles. The ride gap is even larger because Waymo’s vehicles carry paying passengers throughout broad service areas, while many Zoox vehicles are still used for testing, employee transport or limited early access.

Current robotaxi scale, July 2026

Q4Is Zoox growing faster than Waymo?

Zoox may be growing faster in percentage terms, but Waymo is adding far more real activity. Doubling a fleet from 50 to 100 vehicles is not equivalent to doubling a service from 250,000 to 500,000 weekly paid rides.

Waymo reported approximately 50,000 weekly paid trips in May 2024, more than 250,000 in May 2025 and more than 500,000 by early 2026. That is roughly tenfold growth in less than two years.

The more important point is that the absolute additions are accelerating. Waymo added around 200,000 weekly rides during the twelve months to May 2025, then added another 250,000 within roughly ten months.

Zoox has also moved faster lately. It opened Las Vegas, expanded its San Francisco service area, introduced purpose-built vehicles in Austin and Miami, signed a distribution agreement with Uber and increased its production capacity.

App data reported by Business Insider provided one positive Zoox indicator. Zoox reportedly increased its share of active robotaxi-app users from 15% to 25% during the first half of 2026, while Waymo’s share fell from 79% to 69%.

That does not mean Zoox captured one quarter of robotaxi rides. App activity can reflect curiosity, waitlist checks and demand for free rides. Waymo’s growth is visible in completed paid journeys. Zoox’s growth is still more visible in interest, vehicles and launch preparation.

Q5Can Waymo or Zoox open new robotaxi cities faster?

Waymo can currently turn a new city into a functioning robotaxi market much faster than Zoox. Zoox still turns most new cities into testing programs.

Waymo launched fully autonomous public service in Dallas, Houston, San Antonio and Orlando on the same day in February 2026. By May, it was operating across more than 1,400 square miles in 11 cities.

It has also begun driverless employee operations ahead of public launches in cities including San Diego, Las Vegas, Tampa and Denver. The sequence is becoming repeatable: mapping, testing, employee rides, public access and expansion.

Zoox began testing conventional vehicles in Austin and Miami in 2024. Its purpose-built robotaxis arrived later, but ordinary paid service has still not followed at Waymo’s pace.

The difference is operational maturity. Waymo already has procedures for depots, charging, roadside recovery, customer support, remote assistance, local government coordination and vehicle maintenance.

Zoox is building that operating system while validating a new vehicle at the same time. Until its city launches become repeatable, Waymo has the superior expansion model.

Q6Did Zoox build a better robotaxi than Waymo?

Zoox has built the more convincing purpose-designed robotaxi cabin. Waymo has built the better transportation product.

The Zoox vehicle uses no space for a driver’s seat, pedals or steering wheel. Four passengers sit face to face. The vehicle can move in either direction, so it does not need to turn around before beginning its next trip.

Its sensor arrangement is also symmetrical. The vehicle does not have a permanent front or rear, which could reduce low-speed maneuvering and improve curb access in dense areas.

Waymo’s Jaguar vehicles are less elegant. They still carry components and proportions inherited from private cars. The passenger sits in a conventional rear seat inside a large vehicle originally designed for a human driver.

But passengers buy journeys, not architecture. Waymo already provides door-to-door transportation across large parts of multiple cities. Zoox often operates through selected pickup areas, limited operating zones and invitation-based access.

The cabin question goes to Zoox. The complete product question still goes to Waymo.

Q7Can Zoox manufacture enough vehicles to challenge Waymo?

Zoox now has credible manufacturing infrastructure, but it has not demonstrated credible manufacturing volume. Its factory ambitions are measured in thousands of vehicles, while its current fleet is still measured in hundreds.

Zoox’s 220,000-square-foot Hayward facility was designed to support more than 10,000 robotaxis annually at full capacity. The company has also discussed ramping toward 100 completed vehicles per week.

A sustained rate of 100 vehicles per week would equal approximately 5,200 vehicles per year. That would be transformational for Zoox, but it would still take almost seven months merely to match Waymo’s current fleet of around 3,000 vehicles.

Waymo uses a less vertically integrated model. Magna installs Waymo’s autonomous hardware at its Mesa facility, while Hyundai will supply Ioniq 5 vehicles manufactured in Georgia. Waymo has said its integration operations are moving toward annual capacity in the tens of thousands.

Zoox might eventually gain tighter control over maintenance, component integration and vehicle redesign. Today, Waymo’s model is producing more vehicles with less manufacturing risk.

Factory capacity is easy to announce. Factory output is the test.

Q8Does Waymo or Zoox have the stronger safety evidence?

Waymo has the stronger robotaxi safety case by an overwhelming margin because its claims are supported by a much larger real-world sample. Zoox has promising design features, but nothing close to the same exposure data.

Waymo’s June 2026 safety update covered more than 220 million fully autonomous miles. Its analysis found 92% fewer crashes involving serious or fatal injuries than comparable human driving.

Earlier peer-reviewed research examined 56.7 million Waymo driverless miles. It found significant reductions in injury-reported crashes, airbag deployments and several other crash categories.

The human-driving comparison still requires choices about road type, geography and crash reporting. Even so, the scale of Waymo’s dataset makes the results increasingly meaningful.

Zoox has accumulated millions of autonomous test miles, but it has not published a similarly detailed driverless crash comparison across a large public-service fleet.

Zoox emphasizes more than 100 safety innovations in its purpose-built vehicle. Those may improve occupant protection. They do not yet show how the complete system performs after hundreds of millions of miles.

Robotaxi safety evidence, July 2026

Q9Does Waymo or Zoox handle real-world failures better?

Waymo has more experience recovering from unusual failures, but neither company has solved citywide disruptions or emergency scenes. Both still struggle here.

Waymo’s larger fleet has encountered problems that small pilots rarely expose. During major San Francisco congestion around July 4, several vehicles reportedly became immobilized, lost charge or required towing.

A separate power outage disrupted traffic signals and forced Waymo to pause parts of its service. City officials subsequently requested stronger procedures for rerouting, disabled vehicles and emergency coordination.

Zoox faced a different failure after one of its unoccupied vehicles entered a smoke-obscured area near a Las Vegas fire. The vehicle eventually stopped, and Zoox issued an over-the-air software update covering 105 vehicles.

Another Zoox filing described 62 unnecessary lane-crossing events associated with a software issue. The number is specific enough to show that Zoox is identifying repeated patterns rather than treating every event as isolated.

Waymo gets the edge because it has dealt with more incidents across more cities and has a larger recovery operation. The awkward part is that its failures can now affect streets at a scale Zoox has not yet reached.

Q10Which model has the better robotaxi economics, Waymo or Zoox?

Waymo has the more credible economics today because it combines real utilization with declining hardware costs. Zoox has a potentially cheaper end-state vehicle, but no public evidence that it is cheaper per ride.

Waymo’s reported figures imply around 24 rides per vehicle per day. If its average journey is approximately 4.4 miles, that represents roughly 105 passenger miles per vehicle per day before empty repositioning and depot travel.

Those numbers do not prove profitability. The fleet still carries substantial costs for charging, cleaning, maintenance, remote support, depots, insurance and vehicle depreciation.

However, Waymo is spreading those costs across more than half a million paid rides every week. Its sixth-generation Driver is also designed to use fewer sensors and lower-cost components than previous generations.

Zoox has optimized the physical vehicle for utilization. Its long battery life, four passenger seats, bidirectional movement and absence of driver controls should reduce wasted space and inactive time.

But none of the decisive Zoox numbers are public: vehicle production cost, maintenance cost, fleet availability, paid rides per day, empty miles, charging time and revenue per vehicle.

Waymo has imperfect but measurable economics. Zoox has an economic thesis.

Q11Does Waymo or Zoox have the stronger robotaxi distribution?

Waymo has the stronger robotaxi distribution because passengers can already access a functioning service across multiple cities. Zoox has secured an excellent distribution partner, but it still lacks enough available vehicles.

Waymo uses its own app in many markets. In Austin and Atlanta, it distributes rides through Uber. That gives it two acquisition channels: direct customer ownership and access to an established ride-hailing marketplace.

Zoox initially tried to control the whole relationship through its own app. That protects the brand and passenger data, but it makes it harder to fill a new fleet quickly.

Its 2026 Uber agreement addressed that weakness. A dedicated Zoox fleet is expected to appear through Uber in Las Vegas, followed by Los Angeles.

This was a strong strategic decision. Uber can supply demand, payments, route familiarity and millions of existing users.

But distribution cannot compensate for unavailable vehicles. Waymo already has thousands of cars serving passengers. Zoox has an agreement waiting for a fleet.

Q12Is Waymo’s robotaxi data advantage becoming impossible for Zoox to close?

Waymo’s real-world data advantage is widening because it now adds several million autonomous miles every week. Zoox cannot close that gap simply by running more simulations.

At more than four million autonomous miles per week, Waymo can accumulate over 200 million additional miles in one year if its current activity remains stable. That is roughly equal to the entire driverless dataset included in its June 2026 safety update.

Those miles cover different road systems, passenger behavior, construction patterns, weather conditions, emergency scenes and driving cultures across several states.

Zoox certainly has more development data than its public fleet numbers suggest. It has tested retrofitted vehicles for years in San Francisco, Las Vegas, Seattle, Austin and Miami.

But the important gap extends well beyond perception data. Waymo is gathering information about missed pickups, passenger support, vehicle cleaning, charging schedules, road closures, fleet balancing and disabled-vehicle recovery.

That is the dataset required to run a transportation network. Simulation can only cover part of it.

Q13Is Zoox’s unusual robotaxi design slowing it down?

Zoox’s purpose-built design is currently a regulatory disadvantage. The same features that make the vehicle distinctive also prevent it from fitting neatly into rules written for conventional cars.

The Zoox robotaxi has no steering wheel, pedals or traditional driving position. Federal vehicle standards still assume that many vehicles have a human driver, a fixed front direction and conventional controls.

In 2025, NHTSA granted Zoox a demonstration exemption. That permitted research and demonstration activity but did not automatically authorize unrestricted commercial deployment.

Zoox later requested temporary exemptions from parts of eight federal safety standards, including rules related to visibility, lighting and windshield equipment. That process entered public consultation in 2026.

Waymo avoided most of this problem by starting with road-legal production vehicles. Its Jaguars already met the basic structural requirements expected of ordinary passenger cars.

Zoox may eventually help rewrite the rules for autonomous vehicles. Right now, it is waiting for those rules while Waymo collects fares.

Q14Does Waymo or Zoox turn technical progress into commercial service faster?

Waymo converts product progress into passenger service faster. Zoox has recently increased its launch activity, but too many milestones still end at testing, manufacturing readiness or future deployment.

Waymo began driverless operation with its sixth-generation Driver in early 2026. It then introduced passenger rides using the new Ojai platform and continued validating the technology on Hyundai vehicles.

During the same period, it opened several cities, expanded service areas and increased weekly ride volume. The technical system, manufacturing plan and commercial network are moving together.

Zoox also produced a coherent sequence. It expanded Las Vegas and San Francisco, deployed its vehicle in Austin and Miami, opened a larger factory, updated the robotaxi and signed Uber.

The difference appears after the announcement. Waymo’s releases increasingly produce paid rides. Zoox’s releases usually produce another stage of preparation.

Zoox is moving again. It just is not converting activity into commercial output at Waymo’s rate.

Q15Does Amazon give Zoox enough money to catch Waymo?

Amazon gives Zoox enough financial protection to remain in the race, but Waymo currently has the clearer capital advantage. The contest will not be decided by survival alone.

Amazon acquired Zoox for approximately $1.2 billion in 2020. Since then, it has funded autonomous development, the custom vehicle, manufacturing facilities, testing operations and early consumer launches.

Zoox does not disclose standalone losses, annual investment or expected funding requirements. We therefore cannot measure whether Amazon is treating it as a future core platform or a controlled long-term experiment.

Waymo raised $16 billion in early 2026 at a reported post-money valuation of $126 billion. Alphabet remained the majority investor, while external investors supplied additional capital.

That round alone was more than thirteen times Amazon’s original purchase price for Zoox.

Capital buys runway. Waymo has already shown that it can turn that runway into vehicles, cities and paid rides. Zoox still needs to demonstrate the same conversion.

Q16What is the biggest robotaxi bottleneck for Waymo and Zoox now?

City-scale reliability is now the decisive robotaxi bottleneck. Waymo is addressing it more directly because it already operates thousands of vehicles in uncontrolled public environments.

A robotaxi network must handle blocked roads, damaged traffic lights, emergency vehicles, passenger mistakes, airport rules, construction crews, poor weather, dead batteries and thousands of simultaneous routing decisions.

Waymo’s recent failures demonstrate that it has not solved those problems. They also show that it has reached the scale where rare edge cases become daily operating events.

Zoox is spending more of its effort on the vehicle itself: interior efficiency, bidirectional movement, battery duration and sensor placement.

Those improvements may matter later. They do little about the immediate obstacles preventing robotaxis from becoming a dependable urban utility.

The industry already has futuristic cabins. What it lacks is broad service coverage, fast rollouts, regulatory clarity and consistent performance during strange events.

That gives Waymo the strategically more important lead.

Q17Can Zoox still catch Waymo in robotaxis?

Zoox can still become a major robotaxi operator, but design quality alone will not close the gap. Its architecture must produce measurably better operations.

Four results would make the race materially closer.

First, Zoox needs several thousand deployed vehicles rather than roughly one hundred. A sustained production rate of 100 vehicles per week would be an important start.

Second, it needs permission to charge ordinary passengers across multiple cities. Free rides measure interest, not willingness to pay.

Third, Zoox needs to publish utilization and cost data. Its model becomes interesting if each vehicle operates longer, completes more daily rides or costs less per passenger mile than Waymo.

Fourth, the Uber partnership must create repeat usage. Zoox needs tens of thousands of weekly paid rides, not merely app downloads or long waitlists.

The relevant proof would be something like 50,000 paid weekly rides across three cities, supported by at least 1,000 active vehicles and credible safety reporting.

That would still leave Zoox at roughly one tenth of Waymo’s current weekly volume, but it would prove the model can scale. At that point, this becomes a real race.

Q18Who is winning Waymo vs Zoox right now?

Waymo is winning the robotaxi race decisively today. It has the larger fleet, the paid demand, the broader geography, the stronger safety evidence, the faster city rollout and the deeper operating dataset.

The reporting differences cannot explain this gap. Waymo completes more than 500,000 paid rides per week. Zoox does not yet disclose a meaningful paid-ride figure.

Waymo operates around 3,000 vehicles. The most concrete recent Zoox fleet disclosure covers 105 vehicles.

Waymo has analyzed more than 220 million fully autonomous miles. Zoox has not published a comparable driverless safety dataset.

Zoox’s genuine lead is narrower but important: it has designed a more coherent autonomous vehicle. Its cabin, directional flexibility and fleet-oriented architecture may eventually deliver better utilization.

That possibility does not outweigh the current network gap. An elegant vehicle is only one piece of the system. The winner must reliably manufacture, deploy, fill, operate and monetize thousands of vehicles.

Waymo has already built that system. Zoox is still trying to prove that its vehicle can support one.

Waymo vs Zoox robotaxi scorecard, July 2026

Waymo would lose this lead only if its operational problems worsen as the fleet grows, its costs remain structurally too high or regulators slow further expansion.

Zoox can change the answer by turning its architectural advantages into hard numbers: faster manufacturing, more rides per vehicle, lower cost per mile, broader paid service and safety performance across a much larger fleet.

Today, Waymo is winning the network, the business and the operating race. Zoox is winning the argument about what the ideal robotaxi might eventually look like. That is a much smaller prize.