Dara Khosrowshahi has a message for the 9.5 million people who drive for Uber: enjoy the gig while it lasts.
Speaking on the Diary of a CEO podcast — a stage where billionaire-optimism is essentially the house format — Uber's CEO laid out his vision for a future where autonomous vehicles handle most of what Uber does. "You can imagine the majority of our trips being fulfilled by robots of some kind," he said. "Probably not 10 years from now, but you go 15 to 20 years from now, you are going to start getting there."
That's the headline-friendly version. The version that does not fit on a podcast title: Goldman Sachs projects 35,000 robotaxis on U.S. roads by 2030, up from roughly 1,500 today, capturing about 8 percent of the current ride-hail market, per Fortune. That is real growth, but it is not the majority Khosrowshahi is describing.
The gap between the pitch and the current scale is where this story lives.
Uber launched Uber Autonomous Solutions on February 23, a new unit dedicated to commercializing robotaxis globally. The company has signed agreements with roughly 20 AV partners, including Waymo and China's WeRide, and committed more than $100 million to building AV charging infrastructure across the U.S., per Observer. These are not venture statements. They are capital commitments against a specific future. Uber Autonomous Solutions will offer what it describes as an industry-first insurance program alongside real-time roadside assistance.
Waymo is already on the platform. The Alphabet-owned AV company operates in Austin and Atlanta through the Uber app. Khosrowshahi has said Uber expects to facilitate more autonomous rides than anyone else in the world by 2029. The "winning the robotaxi race" framing is not subtle.
Boston Consulting Group estimates that in optimistic scenarios, robotaxis could eventually capture up to 85 percent of trips in large markets, per Observer citing BCG. That is the kind of number that makes the trillion-dollar framing feel almost conservative. But "eventually" is doing a lot of work in that sentence. The 15-to-20-year timeline Khosrowshahi gave Diamandis is not a tech roadmap. It is an admission that the current fleet is measured in the thousands, not millions.
Which raises a question the podcast did not spend much time on: what happens to the people currently filling those seats?
Uber's drivers are not employees. They are contractors, 9.5 million of them worldwide, who own their vehicles, pay for their own fuel or charging, and bear the risk of a market that shifts beneath them. According to Fortune, Khosrowshahi has argued the transition will be generational rather than sudden. That framing is convenient for a company whose current value depends on maintaining a labor model it plans to eventually replace.
Khosrowshahi's own trajectory is instructive. He forfeited roughly $184 million in unvested Expedia stock options when he left for Uber in 2017, and Uber reportedly paid him hundreds of millions to take the CEO job — a figure reported at the time by Wikipedia and Celebrity Net Worth, among others. At the time he took over, Uber was losing $4.5 billion a year — a figure confirmed by contemporaneous reporting from the LA Times and Mercury News. Last year Uber reported net income of roughly $10 billion on $193 billion in full-year 2025 gross bookings, per its investor relations filings.
That turnaround did not happen because Uber built better technology than its competitors. It happened because Uber found a labor model that offloaded cost and risk onto the people driving its cars. The autonomous future Khosrowshahi is selling depends on replacing that model entirely — with capital. Fleets of autonomous vehicles, owned and maintained by third-party operators, with Uber collecting take rates on bookings the same way it does on human-driven trips. Whether the company is selling rides or the infrastructure for others to sell rides, the margins flow upward.
Among Uber's more notable AV partners is Lucid, which it showcased alongside Nuro at CES 2026. TechCrunch reports the partnership calls for up to 20,000 Lucid Gravity robotaxis over six years in dozens of global markets. Uber has invested $300 million in Lucid, per CNBC, and separately made a significantly larger commitment to Nuro, with both tied to U.S. deployment beginning in 2026.
Khosrowshahi has separately said Uber does over $200 billion in bookings on a run-rate basis, per Semafor. Uber reported $193 billion in gross bookings for full-year 2025, per its investor relations filings. The run-rate figure appears to reflect annualized fourth-quarter pace — if Q4 came in around $54 billion, that annualizes to roughly $216 billion, which would explain the spread.
The workers who made Uber's current business possible are not featured in that platform pitch. The podcast with Diamandis was not the venue for a driver transition plan — it was the venue for telling investors that Uber's future is autonomous, capital-intensive, and globally scalable. Khosrowshahi has not offered a plan for the people who built the present.
He told Diamandis that AI will be able to replace the work that 70 to 80 percent of humans can do within 10 years. He said it casually, the way you say something on a stage built for casual audacity. The Goldman Sachs projection — 35,000 robotaxis by 2030 — suggests the replacement of human drivers is not in that first wave. The drivers may have more time than the CEO's broader rhetoric implies. They just do not have a reason to believe it.
† Add verification from a registered source, or add † footnote: "Source-reported; not independently verified."
†† Either cite a registered source that covers this fact, or add verification from one of the registered outlets (e.g., a Fortune or Semafor article from that era), or add † footnote: "Source-reported; not independently verified."
††† Verify against a registered source (e.g., Uber earnings report, Fortune, or Semafor coverage of Uber's financials) or add † footnote: "Source-reported; not independently verified."
