Uber vs. Tesla, Robotaxi Timelines, and the End of Human Driving | Uber CEO Dara Khosrowshahi | #243
Uber CEO Dara Khosrowshahi joins Peter Diamandis (with Salim Ismail asking a couple of questions) for a live fireside chat on the robotaxi race. Dara lays out Uber's hybrid strategy of partnering with over 20 autonomous vehicle makers (Waymo, WeRide, Pony.ai, Nvidia, Lucid, Nuro, Zoox, Wabi) rather than building its own vehicle, an eventual asset-light fleet-ownership model financed by institutional capital, and a plan to convert displaced human drivers into autonomous-fleet owners. He gives concrete but often hedged timelines: 15 autonomous cities by end of 2026, more robotaxi rides than anyone else by 2029, every new car sold autonomous-capable within about 10 years, but a much longer runway before autonomous vehicles reach 50% of the fleet given the 10+ year average car lifespan. The conversation also covers Tesla/Elon's vertically-integrated approach, liability and safety economics, Uber Eats' drone and sidewalk-robot delivery expansion, Middle East and UK leadership in AV/eVTOL adoption, the Uber-Joby eVTOL partnership launching in Abu Dhabi, and Dara's high-intensity management philosophy.
Dara argues autonomy won't be a binary all-or-nothing switch; cities will run mixed fleets of autonomous and human-driven cars, with Uber aggregating demand across both.
Uber has 20+ AV partners (Waymo, WeRide, Pony.ai, Nvidia, Lucid, Nuro, Avride, Wabi, Zoox in discussion) and positions itself as the demand aggregator, data provider, and fleet-management layer across all of them.
Tesla FSD and Elon Musk relationshipTransportation▶ 5:11
Dara says Uber would welcome camera-only FSD Teslas once proven safe without a driver, contrasting Tesla's vertical approach with Uber's multi-partner platform model; outcome with Tesla is described as TBD.
Liability and safety economics of AVsTransportation▶ 10:41
Discussion of how liability shifts from human drivers (Uber's commercial insurance) to AV makers (Waymo owns its own driver's safety case), and how falling liability costs should lower consumer prices.
Dara compares the future AV fleet-ownership structure to Marriott not owning its hotels: Uber will operate/clean/repair the fleets while financial players (e.g., Blackstone-style investors seeking ~9% yield) own the cars.
Timeline for autonomous vehicle adoptionTransportation▶ 15:25
Answering the horse-and-buggy-to-car analogy question, Dara predicts every new car sold will be autonomous-ready within about 10 years, but says the existing fleet's 10+ year average lifespan means full 50% penetration will take much longer.
Peter contrasts Waymo's roughly $150k per-vehicle cost against Tesla Cybercab's targeted $30k, questioning how important asset cost will be to profitability.
Uber Eats and autonomous delivery expansionRobotics▶ 26:53
Dara describes delivery autonomy as harder than passenger autonomy (packages can't move themselves) and outlines a multimodal plan: drone delivery for suburbs, sidewalk robots for short urban trips, and eventual road/bike-lane delivery robots.
International AV and eVTOL leadershipGeopolitics▶ 18:16
Dara names the Middle East (UAE, Saudi Arabia) as the most forward-leaning region for AV and eVTOL adoption, with Europe/UK (partner Wave) also opening up; recalls France's earlier attempt to ban Uber outright.
Uber-Joby eVTOL flying car partnershipTransportation▶ 22:13
Dara recounts Uber Elevate's role catalyzing eVTOL investment before selling the unit to Joby, and describes an end-to-end trip-planning vision (Uber to vertiport, Joby flight, Uber to destination) launching first in Abu Dhabi.
Drivers transitioning from labor to capitalEconomy▶ 24:13
Salim raises Uber's plan to let displaced human drivers become owner-operators of small autonomous fleets, and Dara connects it to Uber's broader strategy of expanding flexible work types (including AI data labeling via Uber AI Solutions).
Dara's management philosophy and cultureOther▶ 30:31
Dara describes Uber's high-intensity work culture, his decision to join a struggling company to have impact, and a new AI 'Dara avatar' employees use to rehearse pitches before meeting him.
Predictions made
openDara Khosrowshahi: Uber will operate autonomous/robotaxi rides across 15 cities with its partners.
“if you think about 10 years from now, I think every single new car sold is going to have autonomous software associated with it”
Your call:
openDara Khosrowshahi: There will be 10-plus different autonomous vehicle providers on city streets (e.g. LA) a decade from now, though the underlying software layer may consolidate to fewer than 10 providers.
“I think it'll be, you know, 10 plus different providers... I don't know if they'll be 10 plus software providers”
Your call:
openDara Khosrowshahi: Within about 10 years every new car sold will be 'autonomous ready' with a full sensor stack (lidar plus cameras), but because the existing US car fleet has an average life of 10-plus years, it will take far longer than 10 years for autonomous vehicles to reach 50% of cars on the road.
EP #? · · due: unspecified, beyond 10 years · ▶ watch
“within 10 years every new car is going to be autonomous ready... it's going to take a very very long time... I think it'll take a while for that fleet to turn over”
Your call:
openDara Khosrowshahi: Uber will have significantly more total drivers on its platform, including in the US, than it has today.
“looking forward to 2030, we're going to have significantly more drivers uh on the platform, including the US, than we have today”
Your call:
openDara Khosrowshahi: Uber and Joby will launch end-to-end eVTOL trip planning in Abu Dhabi, where a rider can push a button, take an Uber to a vertiport, fly a Joby, then take another Uber to their final destination.
“what we're hoping is in Abu Dhabi this year, towards the end of the year, you'll be able to push a button, get an Uber to the vertiport... and then take a Joby”
Your call:
openDara Khosrowshahi: Joby eVTOL service (via the Uber partnership) will expand to the United States.
EP #? · · due: 2027 (sometime next year) · ▶ watch
“hopefully it'll come to the US sometime next year as well”
Your call:
Numbers that matter
Uber was losing $4.5 billion a year when Dara Khosrowshahi joined as CEOPeter recounting Uber's financial turnaround since 2017
Uber now earns over $10 billion a yearPeter contrasting today's profitability with the $4.5B annual loss at Dara's start
Uber has over 20 partners in the autonomous vehicle spaceincluding Waymo, Nvidia, WeRide, Pony.ai, and Wabi
Uber will be in 15 cities with autonomous partners by end of yearAV rollout plan for 2026
Tens of thousands of Teslas are already on Uber's platformdriven by human drivers, some using FSD
Financial players like Blackstone-type investors could target roughly 9% yield owning AV fleetsasset-light fleet-ownership model analogy to Marriott/hotel REITs
Autonomous mobility/delivery described as another trillion-dollar marketplaceDara's framing of the AV opportunity's total addressable market
Tesla Cybercab is targeted to cost around $30,000contrasted against Waymo's ~$150k hardware cost
Uber operates in over 70 countriesnoting many developing markets will be slower to adopt autonomy
An electric autonomous car could be roughly four times cheaper than owning a carPeter citing his own prior research
The French government spent $40 billion trying to fix crime in the Paris banlieues with no impactbefore Uber's entry allegedly changed outcomes
Crime rate in the banlieues reportedly dropped about 10x within 18 months of Uber's entryattributed to residents taking up driving work
Uber has roughly 1 million active drivers in the USand about 10 million globally
Uber's business and audience are each growing at almost 20%context for why Uber still needs to add many drivers despite autonomy
Worth digging into
🕳️ Uber's 2029 robotaxi ride-volume claim
Dara made a bold, specific claim that Uber will facilitate more autonomous/robotaxi rides than anyone else in the world by 2029, which is directly checkable against Waymo and Tesla ride volumes.
🕳️ Waymo vs. Cybercab per-vehicle cost gap ($150k vs $30k)
A 5x hardware cost difference between the two leading robotaxi platforms should heavily determine unit economics and who wins on price.
🕳️ Uber AI Solutions gig-work pivot
Uber quietly repurposing its driver workforce into AI data-labeling and model-evaluation work is a notable and underreported diversification of gig labor into AI training pipelines.
🕳️ Asset-light REIT-style ownership of robotaxi fleets
Dara's Marriott/Blackstone analogy implies a coming financialization of autonomous vehicle fleets as a yield-bearing asset class, which could reshape both AV economics and capital markets.
🕳️ Middle East as an autonomous vehicle and eVTOL leapfrog market
Dara singles out the UAE and Saudi Arabia as the most aggressive adopters of both robotaxis and eVTOLs, potentially outpacing the US and China in real-world deployment.
🕳️ Wave's UK autonomous driving software stack
Wave is described as a UK partner with strong AI talent building a software stack meant to be sold to any and all OEMs, a potentially significant but low-profile player in the AV software layer.