Uber President on Travis, China & Self-Driving | Why Autonomy Is Existential | How to Beat DoorDash

Why does it matter? Because the AV threat to Uber runs exactly backwards from how most people frame it.

Uber's president openly calls autonomous vehicles existential — then explains why that same threat makes Uber structurally harder to displace. Andrew McDonald has spent 14 years inside ride-hailing, knows the unit economics cold, and frames the AV question not as disruption but as distribution. The entity that owns demand doesn't get disintermediated by new supply technology. It gets stronger.

• Waymo and Tesla will likely list their fleets on Uber's network — high-utilization assets need demand, and Uber has it at scale • Autonomous vehicles are already better in some use cases; today is the worst they will ever be • Price — not product quality — is the only real barrier between 200 million and 500 million users • India and Brazil average fares of $2.50–$4.00 mean the majority of Uber trips stay human-driven for decades, even as AV revenue concentrates in US cities

Waymo and Tesla will put their vehicles on Uber's network — because fixed assets need utilization

Waymo and Tesla are probably going to work with Uber. Not because Uber has leverage over them — because expensive fixed assets need utilization. "Whether Whimo or Tesla ends up being the bigger threat, I don't know. I think ultimately it's in both of their interests to put their vehicles on our network."

The model McDonald reaches for is delivery. McDonald's and Starbucks built massive fixed-asset empires with their own apps and direct channels. They still list on DoorDash and Deliveroo, because marketplace demand drives incremental utilization they cannot replicate alone. A fleet of autonomous vehicles follows identical logic: a car sitting idle costs more than a closed restaurant.

The broader frame covers market structure. China already has four or five AV companies competing and the count is rising. The idea that the rest of the world converges on a single winner while China fragments is hard to defend. More competitors means more supply hunting for the highest-distribution demand network. "In the end, distribution wins."

Today is the worst autonomous vehicles will ever be — and Uber's single largest investment reflects that

Every current AV shortcoming — slower pickups, limited weather coverage, imperfect routing — is a today problem, not a fundamental one. "Autonomy is as bad as it's ever going to be today. And every single day it's going to get better."

McDonald doesn't hedge on direction. AV is already a better product in some use cases — not eventually, now. The advantage he emphasizes isn't safety (though he expects AV safety to surpass human). It's the in-car experience: privacy, no obligatory small talk, the ability to work or sleep en route. "People prefer that for the most part." Use cases grow over time; eventually he thinks AV is better in all of them.

The stakes are stated plainly: "If Uber doesn't have autonomy on our platform... it certainly would be existential for our core business." Uber's response is to make autonomy its single largest investment — equity stakes, vehicle purchase commitments, infrastructure build, data collection fleets. Not a hedge. A first-order bet.

Uber One is now Uber's highest-ROI consumer lever — and McDonald spent years actively arguing against building it

The capital went to pricing and driver supply instead. Membership programs meant giving away a suite of benefits when you could put dollars straight back into the transaction. He was wrong. "I probably was short-termist in my thinking there."

The data shifted his view. "I think membership is the most efficient long-term consumer lever that we've got." The mechanism is compounding: members don't just ride more next month — the cohort acquired in any given month keeps riding more over time, folds in Uber Eats benefits, becomes resistant to churn. LTV climbs. At 18 months, the membership ROI looks completely different than it does at 90 days.

The contrast with price promotions is direct. A discount boosts the week you offer it; the signal dissipates fast. Membership moves the other direction. Evaluating both instruments on the same time horizon was the error — and it cost years of compound LTV.

$52 million a week on price subsidies — and DiDi found 200 employees simultaneously on both payrolls

In the final weeks of Uber China, the company was burning $52 million per week — purely on price subsidies, not operations or expansion — as both sides tried to show better economics heading into deal talks. Below that number sit stranger details.

When DiDi merged with rival Kuadi, they found roughly 200 employees appeared on both companies' payrolls at once. Uber, during this same period, had lost access to the WeChat platform. "That's like trying to compete in the US without email or a phone number."

McDonald's conclusion isn't about execution quality: "Even for geopolitical reasons alone, the notion that a US tech company would ultimately be the largest mobility service in China — I just don't think it's something that was ever plausible." There was a ceiling. No amount of subsidy burn was going to break through it. Model the ceiling first, then the addressable market.

AI turned a 15-hour process into 2 hours — and the saved time filled with something else entirely

A capital allocation process that ran 15 hours now runs in 2. A weekly forecasting cycle that took 8 hours now takes 2. Marketing QA dropped from two weeks to two days. The ROI is real and documented at Uber. So where does it go?

"The 8 hours of value that was created gets filled with some other activity which is also presumably high value." Saved time doesn't disappear — employees do more forecasting with more precision, more allocation cycles, more reviews. Output expands. Cost stays flat.

McDonald's prescription: stop trying to connect process improvements to specific headcount reductions. That line is analytically impossible to draw cleanly. Set tighter headcount growth targets and combine compute and people budgets into a single pool. "If we really believe that AI is making our employees 20% more efficient, then next year we should just not increase headcount." That's the extraction mechanism. Everything else is a rationalization.

Uber X at $35 a direction is still a luxury product — and the next 300 million users can't afford it

Thirty-five dollars each way in New York. McDonald calls that a luxury product outright. "The vast majority of transactions in transportation broadly happen at a price point that is way lower than our core products." Getting to 500 million users — or from 6 average monthly trips to 25 — requires transaction costs to fall structurally, not incrementally.

This reframes autonomous vehicles entirely. The in-car privacy and no-awkward-driver experience is a premium for current users. The real AV value to Uber is cost compression: remove the driver, bring the fare down, reach the next 300 million people who cannot afford daily Uber use today.

Brazil averages $3.50–$4.00 per trip. India averages $2.50–$3.00. Both are top-three Uber markets by volume. AV economics won't reach those price points for decades. The wave hits US cities first — also where revenue concentrates. Measure AV impact by revenue share in premium urban markets. Global trip share is a distraction.

The AV companies that scare investors most are also the ones that need Uber most

What McDonald describes almost offhandedly is a structural advantage most disruption narratives miss. Uber doesn't compete with AV technology — it benefits from every AV company that scales, because more autonomous supply means more options flowing through a demand network that compounds with every new user. The scarier the autonomous threat looks from outside, the more every AV company needs Uber's distribution. Physical-world technology will keep improving. The entity that owns the relationship with 200 million monthly users collects a toll on all of it.

---

Topics: Uber, autonomous vehicles, ride-sharing, membership programs, AI ROI, China market exit, distribution strategy, Travis Kalanick, Dara Khosrowshahi, food delivery, DoorDash, Waymo, Tesla, enterprise AI, marketplace economics