Waymo's robotaxi business is chasing one number in 2026: one million paid rides per week in the US by year's end, up from roughly 400,000 today across six cities. Co-CEO Tekedra Mawakana has called the milestone an "inflection point." At about $20 per ride, one million weekly rides works out to roughly $1 billion in annualized revenue — a real business, and a small fraction of what Waymo raised in a single quarter this year.
That gap is the story. Autonomous ride-hailing is finally producing revenue at a scale worth measuring, and the unit economics still do not obviously work, for a reason that gets glossed over constantly: removing the driver does not remove the labor.
This article works through Waymo's actual numbers, what "no driver" costs in practice, how the expansion to 20+ cities including London and Tokyo changes the math, and where the competition stands.
Key Takeaways
- Waymo targets 1 million paid weekly US rides by end of 2026, up from roughly 400,000 per week across six cities.
- At about $20 per ride, that is roughly $1 billion in annualized revenue — against a $16 billion raise in Q1 2026 alone.
- Driver labor is 60-70% of traditional ride-hailing cost, but robotaxis add remote assistance, depot operations, cleaning, and hardware amortization.
- Expansion plans cover 20+ cities in 2026 including first international launches in London and Tokyo.
- The new Ojai vehicle is roomier and cheaper to build — vehicle cost, not software, is the binding constraint on scale.
Where Waymo actually stands in 2026
Waymo is running roughly 400,000 paid rides per week across six cities, with a stated goal of one million weekly by the end of 2026 and an expansion to more than 20 cities including London and Tokyo. That is a 2.5x growth target in under a year, on top of a physical fleet that has to be built, shipped, mapped, and maintained.
The trajectory is steep and it is real. TechCrunch's ridership analysis charted the curve earlier this year, and the shape is the one you want if you are Alphabet: consistently compounding, not a step function driven by a single city launch.
The international expansion is the more interesting commitment. London and Tokyo are not easier markets than Phoenix — they are dense, heavily regulated, and in Tokyo's case involve left-hand traffic, unfamiliar road markings, and a regulatory culture that does not move quickly. Committing to both in the same year signals that Waymo believes its perception stack generalizes across driving cultures, which is a much stronger claim than "it works in Chandler, Arizona."
Vehicle supply is the quiet constraint. Alphabet is adding the newly designed Ojai to the fleet — roomier and, importantly, cheaper to build than the retrofitted Jaguars it is supplementing. Every robotaxi analysis eventually collides with the fact that this is a hardware business wearing a software business's valuation.
Does removing the driver actually remove the cost?
Partially — and far less than the headline 60-70% figure implies. Driver compensation genuinely is 60-70% of traditional ride-hailing operating cost, and Waymo does not pay it. But autonomy substitutes a different labor stack rather than eliminating labor.
Here is what a robotaxi fleet pays for that an Uber does not:
| Cost line | Human ride-hailing | Robotaxi |
|---|---|---|
| Driver compensation | 60-70% of opex | $0 |
| Vehicle capital | Driver's own car | Fleet-owned, ~$100k+ per unit with sensors |
| Sensor maintenance & calibration | None | Ongoing, specialized |
| Remote assistance staff | None | 24/7, scales with fleet |
| Depot operations, cleaning, charging | Driver absorbs | Fleet absorbs |
| Mapping & city onboarding | None | Large upfront cost per city |
| Insurance | Driver-carried | Fleet-carried, novel risk class |
The line that gets least attention is remote assistance. Autonomous fleets maintain human operators who resolve situations the vehicle will not handle alone — an unusual construction zone, a police officer waving traffic through a red light, a blocked lane with no legal path around it. That is a real payroll that grows with fleet size, even if the ratio of humans to vehicles is dramatically better than one to one.
The cleaning and charging burden is similarly underrated. A human driver notices someone spilled a coffee and deals with it because they have to sit in the car. A robotaxi requires a depot, a schedule, and staff. Multiply by a fleet targeting a million rides a week.
None of this means the economics fail. It means the honest question is not "does removing the driver save 65%" but "does removing the driver save more than the fleet stack costs" — and at sufficient utilization, it plausibly does. Utilization is the whole ballgame, which is exactly why Waymo is chasing a volume number rather than a margin number.
Is $1 billion in annualized revenue impressive?
In isolation, yes — it makes Waymo the first autonomous vehicle company with a revenue line worth putting on a slide. In context, it is modest relative to what the business consumed to get there.
Waymo raised $16 billion in Q1 2026. Against that, roughly $1 billion in annualized revenue at the year-end target means the company would need many years at the target run rate simply to return a single quarter's funding — before considering the cost of generating that revenue.
That framing is not a criticism so much as a description of what stage this industry is at. Capital is doing the work that revenue eventually has to. Waymo was one of four companies — with OpenAI, Anthropic, and xAI — that together absorbed nearly 65% of global venture investment in Q1 2026, a concentration Crunchbase documented as the most extreme in venture history.
The comparison that matters is the trend, not the level. If ride volume grows 2.5x while cost per ride falls — cheaper vehicles, better utilization, fewer remote interventions per mile — the curves cross. If volume grows and per-ride cost stays flat, more rides just means losing money faster. Waymo has not published enough to tell which is happening, and that opacity is deliberate.
How does the competition compare?
Waymo leads on paid-ride volume and city count. Tesla leads on fleet size potential and cost per vehicle. Chinese operators lead on regulatory speed at home. Nobody leads on profitability.
- Tesla is expanding robotaxi service across seven US metropolitan areas, with roughly 700,000 paid miles logged since its June 2025 launch. Note the unit: miles, not rides. Miles are a much easier number to make large, and the substitution in reporting is worth watching for.
- Uber is assembling a partner network rather than building a stack — a Munich pilot planned for late 2026 pending regulatory approval, and a WeRide partnership extending to 15 more cities by 2030. Uber's bet is that demand aggregation outlasts any single autonomy vendor.
- Zoox demonstrated a purpose-built vehicle at CES 2026 and expanded into San Francisco.
- Momenta is preparing a luxury robotaxi service in Abu Dhabi with Mercedes-Benz and UAE operator Lumo.
The industry-wide projection is a global fleet roughly ten times its current size by 2030, meaning more than 100,000 driverless taxis in service. That is large in absolute terms and small against global ride-hailing, which is the right way to hold it: robotaxis are becoming a real category, not yet a replacement for the incumbent one.
The underlying enabler across all of these is onboard compute. Waymo vehicles run server-grade CPUs, GPUs, and specialized accelerators processing camera, lidar, and radar streams in real time — the same silicon economics reshaping data centers, which we covered in custom AI chips breaking Nvidia's grip. Every watt saved on the inference stack is range, and range is utilization.
What robotaxis tell you about embodied AI
Robotaxis are the most demanding deployed test of physical AI, and the pattern that shows up here shows up everywhere embodied systems meet the real world: perception is solved enough to ship, and the long tail of edge cases is what determines cost.
That same pattern is visible in industrial and humanoid robotics, where 2026 was the year deployments stopped being demos — we counted the actual numbers in humanoid robots deployed in 2026. It is visible in general-purpose robot control models too, like the whole-body systems we covered in Gemini Robotics 2.
The lesson generalizes cleanly: in embodied AI, the last 2% of scenarios consumes most of the operating budget. A robotaxi that handles 98% of situations autonomously still needs a human available for the other 2%, and that availability is a fixed cost you pay whether or not it gets used.
Frequently asked questions
How many rides does Waymo do per week?
Waymo provides roughly 400,000 paid rides per week across six cities as of 2026, with a stated target of one million paid weekly rides in the US by the end of the year. That target would put annualized revenue at approximately $1 billion.
Is Waymo profitable?
Waymo has not disclosed profitability, and the available figures suggest it is not. Roughly $1 billion in annualized revenue at the year-end target sits against $16 billion raised in Q1 2026 alone, and robotaxi fleets carry substantial costs that human ride-hailing does not.
How much does a Waymo ride cost?
Estimates put average revenue at approximately $20 or more per ride, broadly comparable to premium human ride-hailing in the same markets. Waymo has not published a full fare schedule, and pricing varies by city and demand.
Which cities will Waymo operate in?
Waymo currently operates in six cities and plans to expand to more than 20 during 2026, including its first international launches in London and Tokyo. The international expansion is the more technically demanding step, given different traffic conventions and regulatory regimes.
Do robotaxis eliminate labor costs?
Not entirely. Driver compensation, which is 60-70% of traditional ride-hailing operating cost, goes to zero — but robotaxi fleets pay for 24/7 remote assistance operators, depot staff, cleaning, charging, sensor maintenance, and per-city mapping. Autonomy substitutes a different labor stack rather than removing labor.
Who is winning the robotaxi race?
Waymo leads on paid ride volume and commercial city deployments. Tesla is expanding across seven US metros and reports mileage rather than ride counts. Uber is aggregating third-party autonomy partners instead of building its own. No operator has demonstrated profitability at scale.
The verdict
Waymo is the clear leader in autonomous ride-hailing and one million weekly rides would be a genuine milestone — the first time this technology produced revenue at a scale that shows up in a corporate filing rather than a press release.
But hold the "60-70% cost saving" framing at arm's length. Removing the driver removes the largest single line item and adds five smaller ones, and whether the trade is positive depends entirely on utilization. Waymo chasing volume rather than margin tells you which side of that equation the company is still working on.
The technology question was settled a while ago. The one that decides this industry is whether a car that drives itself can also pay for itself.
Top comments (0)