Uber Technologies has spent years positioning itself as the platform that could one day integrate fleets of self-driving cars. The promise always sounded straightforward. Remove the driver. Lower costs. Expand margins. Yet recent moves by major investors tell a different story. Polen Capital Management sold its entire stake in the ride-hailing giant during the second quarter of 2026. The reason cited in its investor letter struck at the heart of industry anxiety.
“While Uber is growing nicely, it is difficult to disprove the negative narrative on autonomous vehicle disruption so we decided to use the proceeds elsewhere,” the firm wrote, according to a Yahoo Finance report.
That single decision reflects broader doubts. Uber’s human-driven model still generates the bulk of its revenue. But competitors such as Alphabet’s Waymo and Tesla are building direct robotaxi services that could sidestep Uber’s network entirely. The risk isn’t theoretical. It’s already showing up in headlines, regulatory scrutiny and shifting market sentiment.
Consider the data emerging from real-world deployments. A report from Open Plans and the Sam Schwartz Transportation Program claimed Waymo vehicles in New York City posted a significantly higher rate of crashes resulting in serious injury or fatality compared with the existing for-hire fleet that includes Uber and Lyft cars. The Washington Post examined those numbers closely. It found the report’s own data contradicted the headline assertion. Waymos, the analysis showed, were involved in far fewer injury-causing crashes than either New York City drivers or for-hire vehicles.
Such conflicting signals complicate the narrative. Proponents point to overall reductions in crashes tied to human error. Federal figures have long shown that roughly 94 percent of accidents stem from driver mistakes. Yet autonomous systems introduce new failure modes. Software glitches. Edge cases in perception. Brittle performance when conditions stray from training data. A Brookings Institution analysis from 2024 highlighted how computer vision can misread signs or fail to classify articulated vehicles correctly. Those errors don’t happen often. When they do, they create memorable incidents that fuel public skepticism.
Uber itself carries historical baggage here. The 2018 fatal crash in Tempe, Arizona, where an Uber test vehicle struck and killed Elaine Herzberg, still shadows discussions. The National Transportation Safety Board faulted both the company’s safety culture and the vehicle’s inability to classify the pedestrian. The backup driver faced charges. Years of litigation followed. Even though the operator ultimately received probation rather than prison time, the episode demonstrated how one tragedy can reshape corporate strategy. Uber pulled back from full self-driving development and pivoted toward partnerships.
Today that approach carries fresh complications. Uber has teamed with companies like Motional and Avride to offer autonomous rides through its app. Those alliances bring scale. They also spread risk. In May 2026 the National Highway Traffic Safety Administration opened an investigation into Avride after more than a dozen crashes, including one minor injury. Many occurred in Dallas and Austin, Texas, where Uber had recently launched robotaxi service with the partner. The regulator cited issues with lane changes, responses to other vehicles and stationary objects. TechCrunch reported that at least one crash involved a passenger-carrying vehicle.
And. This matters. Because any incident tied to an Uber-branded autonomous ride can tarnish the entire platform. Riders don’t always distinguish between Uber’s own technology and that of a third-party partner. Public trust erodes quickly when videos of errant robotaxis circulate on social media.
Wall Street has taken notice. A Bloomberg analysis ahead of Uber’s earnings earlier this year captured the tension. “While the core business trends remain strong, autonomous vehicle headline risk continues to dominate discussion around Uber and creates ongoing volatility,” wrote JPMorgan analyst Douglas Anmuth. The February 2026 piece noted that threats from Waymo and Tesla were likely to overshadow operational results. Shares have swung on such news. Uber closed near $75 in mid-August 2026 with a market value above $150 billion. Yet the stock has traded in a wide range over the past year.
Recent X conversations echo the uncertainty. Waymo co-CEO Dmitri Dolgov argued that cameras alone cannot deliver the safety levels required for full autonomy. Tesla, of course, bets heavily on vision. Uber continues to cooperate with Waymo despite occasional rumors of tension. One post from investor Martin Varsavsky captured the mood. “Markets work better when competing technical bets can reach customers instead of being settled in conference rooms.” The founder noted that trust forms in edge cases, not smooth demonstrations.
Safety statistics offer some reassurance. A Johns Hopkins Public Health analysis from July 2026 reviewed early data and found driverless cars can reduce serious crashes. Experts remain cautious about broad claims. Different road users face different risks. Pedestrians, cyclists and other drivers all interact with these systems in ways that are still being measured. The University of Michigan’s Center for Sustainable Systems factsheet notes there simply isn’t enough mileage yet to declare autonomous vehicles definitively safer than humans across all conditions.
Public opinion adds pressure. A poll released by the Advocates for Highway and Auto Safety in April 2026 found 81 percent of respondents concerned about sharing roads with driverless cars. More than 40 percent described themselves as very concerned. Strong majorities backed new federal requirements. Vision tests for the vehicles. Clear disclosure of operational domains. Mandatory crash reporting. Standards for remote operators who sometimes assist from overseas locations.
Senator Ed Markey has pushed on the remote assistance angle. His office released findings showing gaps in transparency and training for those human backups. The issue gained attention after revelations that some Waymo remote staff operate from the Philippines. Such arrangements raise questions about response times, cultural familiarity with local roads and accountability when things go wrong.
Uber’s own autonomous mobility page promotes tools for fleet efficiency, remote assistance and specialized insurance. The company positions itself as the infrastructure layer that manufacturers and operators need. That strategy buys time. It doesn’t eliminate the longer-term threat. If Waymo or Tesla build closed ecosystems that connect directly to riders, Uber’s matching network could lose relevance. A Road to Autonomy analysis described this as Uber’s autonomy paradox. Its massive demand network is both strength and vulnerability.
Academic researchers see parallel tensions. A UCLA Anderson working paper examined the tricky transition period. As autonomous vehicles roll out gradually, they can boost profitability by eliminating driver pay. Yet they risk alienating the very workforce that built the platform. Poor service during the hybrid phase could drive customers away. Lower driver earnings might reduce supply. The paper’s authors, including Francisco Castro, warned that mismanaging this handoff could hurt the business rather than help it.
Regulatory momentum is building toward 2026 as a pivotal year. The National Transportation Safety Board has called for stronger standards on automated driving systems. Aurora Innovation argued in January that federal rules are overdue. The company pointed to simulations showing its technology would have prevented dozens of fatal highway crashes. Human error remains the dominant cause of roadway deaths. Nearly 40,000 Americans lose their lives each year in traffic incidents. That grim total has barely declined in decades.
Yet the path to widespread adoption is anything but smooth. Weather. Construction. Unpredictable human behavior. These factors still challenge even the most advanced systems. A 2026 update on companies like Zoox and Waymo noted promising tests in controlled cities. It also highlighted a recent incident where a Waymo vehicle struck a child near a school. Such events, however rare, reset the conversation.
Uber finds itself caught between two imperatives. It must defend its current driver-powered model while preparing for a future that could render parts of that model obsolete. Partnerships provide a hedge. They also expose the company to partners’ mistakes. Insurance programs tailored for autonomous operations represent an innovation. So do depot management tools and real-time monitoring. These capabilities could position Uber as an essential operator even if it doesn’t own the underlying technology.
But the negative narrative persists. Hedge funds still hold Uber shares in large numbers. It ranked high on lists of popular stocks at the start of 2026. Sentiment can shift fast when headlines turn sour. One serious crash involving an Uber-partnered robotaxi could trigger regulatory pauses, lawsuits and loss of rider confidence.
So the question lingers. Can Uber thread the needle? Or will autonomous driving ultimately erode the very advantages that made the company dominant? Investors like those at Polen Capital have already placed their bets. Others are watching closely as more miles accumulate and regulators weigh new rules. The data will decide. Not the promises.