2026-04-02

Uber’s Robotaxi Playbook, End of Human Driving & $10B Bet on Robots | Dara Khosrowshahi (Uber CEO)

Live audience Q&A with Uber CEO Dara Khosrowshahi, hosted by Peter Diamandis, including questions from a group of teens. Dara says riders accept autonomous vehicles easily (80% say yes when matched), that human drivers will be demonstrably less safe than AVs within about 25 years, and that kids now 13 will likely not need to drive by about 2036. He argues automation mostly augments work, targets 20 million people on Uber's platform by 2035, and expects few combustion engines in 20 years. He also covers insurance layers for AVs, working with Chinese AV and EV partners, and why big companies should take bigger bets.

▶ Watch on YouTube

Topics

Will human driving be phased out? Transportation ▶ 0:00

Teen asks if human driving becomes illegal and what happens to race car driving. Dara says not within 15 years, but within 25 years humans will be demonstrably less safe than AVs, and regulators will decide what a license looks like.

Rider acceptance of AVs Transportation ▶ 1:01

In Atlanta and Austin, riders are told they were matched with an autonomous vehicle: 80% accept, 20% decline, and the 80% love it. Regulators move slower than riders.

Vertiports and flying cars (Joby) Transportation ▶ 3:33

Real-estate question on air taxis. Vertiports must handle many vehicles, will sit at airports and high-traffic city centers, and Uber has data on the best sites.

Marketplace building: supply first Economy ▶ 4:44

Uber marketplaces are supply-led: build liquid supply with real tools and product-market fit, and demand shows up. Contrasted with Expedia, which focused on audience.

Insurance and liability for AVs Economy ▶ 5:55

Dara expects layered insurance: the AV provider insures the vehicle and driver quality, with another layer on top. About half of US court cases are car-accident related.

Dispatch at scale and multi-robot coordination Robotics ▶ 7:29

Uber predicts 5-6 seconds ahead when dispatching and manages 10 million drivers and 40 million trips a day. Machines are more predictable and will accept rides more often than humans.

Universal basic services and access Economy ▶ 10:41

Mark pitches a $250/month affordable-living moonshot (XPRIZE Visioneering winner). Dara commits to serving outer boroughs and transportation deserts and says AVs will eventually make transport cheaper.

China, Chinese AV players and emerging markets Geopolitics ▶ 13:15

Uber exited China before Dara joined and he agrees with that. It holds a stake in Didi and partners with BYD, Pony, WeRide and hopefully Baidu, building bridges to China.

AV insurance as an actuarial problem Economy ▶ 15:44

Insurance model is undecided (driver, platform or both). AV trips were under 1% of Uber's volume growth last year, so there is not yet enough volume for actuaries.

Do teens still need a driver's license? Transportation ▶ 17:20

Dara tells a 13-year-old to still get a license: AVs are expensive today, but in 10 years you will not have to drive. Peter predicts his own son will not need one in two years.

AI, robotics and jobs Economy ▶ 18:52

Dara says automation historically augments work, citing Chinese factories where humans supervise robots. Uber aims for 20 million people on its platform by 2035, with retraining and labor ownership as public and private tasks.

EVs replacing combustion engines Energy ▶ 24:03

Physical-world transitions are slower than digital ones and US charging infrastructure lags. Because AVs are EVs, Uber's network will shift to EVs, and in 20 years few combustion engines remain.

Culture: do the right thing, take big bets Other ▶ 25:36

Core value is 'do the right thing, period', meaning personal responsibility at every level. Larger companies should take bigger bets, since Uber now cash flows $10B a year versus losing $4B when Dara joined.

Entering new fields: it must rhyme Other ▶ 29:20

Uber expands into adjacencies where it has a core skill and right to win: rides, Eats, Freight, Elevate, and Uber AI Solutions. If it does not rhyme with what Uber does, a startup should do it.

Predictions made

open Dara Khosrowshahi: Human driving will not be meaningfully displaced within 15 years, but within 25 years (about 2051) humans will be demonstrably less safe than autonomous drivers.

· due: 2051 · ▶ watch

“in the next 25 years, humans will be demonstrably less safe than autonomous drivers.”

Your call:

open Dara Khosrowshahi: There will be fewer and fewer human drivers on the road over time, and regulators will decide what a human license looks like, possibly with a much more demanding driving test.

· due: unspecified (long term, 25 years) · ▶ watch

“there're going to be fewer and fewer drivers on the road. Just like there are less people who know how to ride a horse.”

Your call:

open Dara Khosrowshahi: A 13-year-old will still want a driver's license at 16 (about 2029) because mass AV production takes time, but 10 years from now (about 2036) people will not have to drive and will have plenty of choices.

· due: 2036 · ▶ watch

“10 years from now um you won't have to drive and you'll have plenty of choices.”

Your call:

open Peter Diamandis: In two years (about 2028) his son will not need a driver's license, and the need for one will collapse over the next few years.

· due: 2028 · ▶ watch

“I predict today that in two years he won't need to get a driver's license.”

Your call:

open Dara Khosrowshahi: By 2035 Uber will have 20 million people earning on its platform, up from about 10 million now, possibly doing different kinds of tasks.

· due: 2035 · ▶ watch

“in 2035, we've got 20 million people on our platform versus the 10 million people on our platform now.”

Your call:

open Dara Khosrowshahi: Over the next 10 years (to about 2036) there will be a lot of augmentation of work and some replacement of work, with new work showing up as society adjusts.

· due: 2036 · ▶ watch

“over the next 10 years, you will see a lot of augmentation of work um and some replacement of work.”

Your call:

open Dara Khosrowshahi: In 20 years (about 2046) there will not be too many combustion engines on the road, as the Uber network moves to EVs along with autonomy.

· due: 2046 · ▶ watch

“I think 20 years from now, yeah. I don't think you're going to see too many combustion engines”

Your call:

open Dara Khosrowshahi: Autonomous vehicles will first be safer and then eventually make transportation cheaper, available to everyone regardless of where they live.

· due: unspecified (eventually) · ▶ watch

“eventually autonomous will make transportation cheaper as well.”

Your call:

open Dara Khosrowshahi: AV insurance will settle into layers: AV provider insures the vehicle and driver quality, with Uber covering another layer on everything else.

· due: unspecified · ▶ watch

“you're going to have different layers of insurance as well.”

Your call:

open Dara Khosrowshahi: Some of Uber's big bets will fail, but Uber will keep making them because at $10B cash flow it can absorb failures.

· due: unspecified · ▶ watch

“some of them are going to fail.”

Your call:

Numbers that matter

80% accept / 20% decline Share of Uber riders in Atlanta and Austin who say yes when matched with an autonomous vehicle.
15 and 25 years Dara's windows: not within 15 years, but within 25 humans are demonstrably less safe than AVs.
About 50% of US court cases Share of legal court cases in the US that are car-accident related, per Dara.
9-10 million drivers Uber's drivers globally.
40 million trips per day Uber's daily trip volume across markets.
$250 a month Mark's affordable-living-as-a-service target for housing, food, energy, connectivity and mobility; prize is $50 million.
70% of cost Chinese AV makers' estimate of cost removed by eliminating the driver (Charles, audience).
70 countries Uber's footprint as stated by the questioner.
Under 1% of volume growth All autonomous trips worldwide last year versus the rides Uber added in the year.
10 years Dara's timeline until people won't have to drive.
20M vs 10M on platform by 2035 Uber's target for people working on the platform versus today.
20 years Dara's timeframe for few combustion engines remaining.
-$4B to +$10B cash flow Uber was losing $4 billion when Dara joined; it will cash flow $10 billion this year.

Worth digging into

🕳️ Uber's AV partner stack and where each robotaxi launches

Dara names Atlanta and Austin AV matching, Chinese partners Pony/WeRide/Baidu and BYD, and a stake in Didi; these determine who wins the ride network layer.

🕳️ AV insurance and liability models

Dara says the model is unsettled (driver vs platform vs both) and that about half of US court cases are car-accident related, implying a huge liability shift.

🕳️ Checking the 80% AV acceptance rate

Uber's claim that 80% accept an AV match and love it underpins demand for robotaxis.

🕳️ 20 million people on Uber's platform by 2035

A concrete labor-platform target that implies Uber expects AVs and humans to coexist and new task types (e.g. AI data labeling) to grow.

🕳️ Peter's prior prediction: son never goes to university

Ties to the show's wider view on degrees and education in an AI era, repeated with the 'no driver's license in two years' prediction.

🕳️ Physical-world adoption lag and EV charging

Dara argues atoms transitions are slower than bits and that AVs drag EV adoption forward.