Uber Says 700 Driverless Vehicles by Year-End and 2,000 in 2027: Did It Promise “Tens of Thousands” by 2028?

Uber gave its first driverless fleet targets on November 4, 2026, during its Q3 earnings call: about 700 fully driverless vehicles on its network by year-end and more than 2,000 by the end of 2027. Headlines then jumped to 2028, saying Uber aims for "tens of thousands" of robotaxis. The transcript line is narrower. Uber's CFO said that as the company scales to "the 10 thousand+ in 2028 and beyond," it wants third-party capital to pay for it. This post separates what Uber said from how it was summarized.

TL;DR: claimed versus verified

What Uber said

According to a Motley Fool transcript of the call, Khosrowshahi said: "In AVs, our focus is shifting from partnerships and testing driverless commercial service." He said Zoox will launch on Uber in Las Vegas later this month, with Motional moving up to fully driverless operations soon after. He then gave the targets: about 700 fully driverless vehicles by year-end and more than 2,000 by the end of 2027. And, he said, "we are focused on demonstrating utilization and economics that are needed to scale." Uber's own Q3 2026 results release is the primary source for the quarter's figures; we could read the transcript text but not the full release text, so we cite the release only as the primary source page.

The financing answer is where 2028 comes from. The CFO said the long-run expectation is that fleets are financed by third-party capital, not Uber's balance sheet. In the near term, with 700 vehicles this year and 2,000 by the end of next, "the capital deployment against those is fairly limited." Then: "as we start scaling out into the 10 thousand+ in 2028 and beyond, we want to bring in third party capital again." That is a statement about how scale would be paid for if it comes. It is not a promise with a date and a unit count.

Why the wording matters

The Motley Fool's own summary says Uber plans to scale its driverless fleet to tens of thousands of vehicles by 2028. A TechTimes headline says "10,000 by 2028." The spoken line is "10 thousand+." Those are three different claims. Ten thousand plus is a floor, tens of thousands is a range several times higher, and a floor in a financing answer is not a commitment. If you quote a number in a story or an investment note, quote the transcript.

The numbers also look small beside Uber's scale. The Motley Fool summary reports 215 million monthly active platform consumers. A fleet of 700 vehicles is a rounding error on that base, which is why the CEO talks about proving utilization and economics first. The question is whether an autonomous vehicle earns more per hour on Uber's network than it costs to own and run, and that is something only operating data will answer.

The financing idea: a utilization premium

The most concrete business claim on the call is about utilization. Khosrowshahi said vehicles on the Uber network see about a 30% utilization premium compared with vehicles off the network, per the transcript summary. He argued this lets Uber move a vehicle flexibly: if a fleet operator fails, the car can be placed under separate ownership and still earn on Uber's network. From that, he said, financing on the Uber network "can have a real advantage in terms of cost of capital versus alternatives."

He added that the model still has to be established. Partners are "quite interested," in his words, because they want to diversify away from their current financing sources, but Uber has to demonstrate the utilization advantage to banks first. TechTimes reports that Uber described a $1 billion Santander arrangement as an early template, and that management noted residual values of autonomous vehicles remain unproven. We could not confirm the Santander detail on the call transcript, so treat it as that outlet's report. The 30% figure is Uber's own number, drawn from its network, and no outside party has audited it.

This is why the 2028 line sits inside a financing answer. A fleet of 10,000 or more vehicles costs far more than Uber wants on its balance sheet. The plan only works if lenders accept that autonomous vehicles keep their value and stay busy, and nobody has years of data to prove either point.

The quarter in numbers

TechTimes, reading Uber's release, reports Q3 Gross Bookings of $59.8 billion, up 20% in constant currency, the fifth straight quarter above that mark. It reports non-GAAP operating income up 38% to $2.3 billion, adjusted EBITDA up 31% to $2.96 billion, and non-GAAP EPS of $0.88, up 35%. The transcript summary reports 4.1 billion trips, up 16%, and 215 million monthly active platform consumers. GAAP net income was $512 million, reduced by a $476 million pre-tax loss on the revaluation of equity investments, per TechTimes.

TechTimes also reports that fully driverless rides are still below 0.5% of Uber's volume and that Uber's partner list exceeds 25 autonomous developers, including Waymo, Zoox, Wayve, Baidu, Rivian and Nvidia. It lists planned deployments: Zoox in Las Vegas in November and Los Angeles in mid-2027, a Wayve and Nissan pilot in Tokyo, and an Nvidia-based fleet starting in Los Angeles and San Francisco in the first half of 2027. Those details come from that outlet's reading of earlier announcements and are not all on the call transcript, so check each partner's own statement before relying on dates.

Note the contrast inside one article: TechTimes' headline says 10,000 by 2028, and its body says "tens of thousands by 2028." The two lines do not match, and neither matches the CFO's spoken "10 thousand+." This is a clear example of how a number drifts as it is retold.

How it compares with other robotaxi plans

Other operators have bigger public numbers. See our reports on Waymo's fleet plans of 20,000 vehicles by 2027 and its expansion to 14 US cities, and on Bolt and Lucid's 25,000-robotaxi plan with Nvidia. Uber's role is different: it is a marketplace that plugs partners' vehicles in. Our post on Uber and Wayve in the UK shows the same model abroad. A marketplace needs partners who can deliver vehicles, and that is where the 2028 number depends on others, not Uber.

Why this belongs in an AI blog

Robotaxis are one of the few places where AI driving models meet a paying public at scale. The partners Uber lists, such as Wayve with an end-to-end learned driver and Nvidia with a reasoning-based driving model, bet that learned systems can generalize across cities without heavy mapping. A target of 700 vehicles is small, but it is the first point where that bet gets measured in trips and revenue per vehicle, not demos. TechTimes' write-up of the release frames the targets as the first time Uber tied a fleet number to a calendar date, which is fair, and which also means the year-end count is easy to check in about eight weeks. If it lands near 700, the 2,000 target gets more credible. If it slips, the 2028 numbers deserve extra doubt.

Other things said on the call

• The CFO named elevated fuel prices in markets such as Brazil and India as a drag on international results.
• The CEO said there will be more than 10 AI assistants through which consumers can reach Uber. We covered the first numbers on that in Uber's four rides from third-party AI agents.
• The company is integrating acquisitions including EZCater and Blacklane.

What to watch next

1. Whether Zoox launches in Las Vegas on schedule later in November.
2. Whether Motional reaches fully driverless operation soon after.
3. Whether the year-end count of about 700 holds. It is the nearest and easiest target to check.
4. Whether Uber publishes utilization and unit economics for driverless vehicles.
5. Whether a 2028 target appears in a formal filing or investor deck, as opposed to a call remark.

How to use this if you follow the sector

Treat the 700 and 2,000 as the actual targets. Treat 2028 as an aspiration tied to financing. When a headline gives a bigger number, find the sentence it comes from. This is the same habit we use in all our claimed-versus-verified posts.

Related reading

• Waymo 20,000-robotaxi fleet report
• Waymo 14 US cities
• Bolt and Lucid 25,000 robotaxis
• Uber and Wayve UK
• Uber's rides from third-party AI agents

Quotes come from a published call transcript read on November 4, 2026. Follow @explainx_ai for updates.