In April 2025, an 18-wheeler pulled onto Interstate 45 outside Dallas with no one in the cab. It merged into traffic, held its lane for roughly 200 miles, and arrived at a freight hub near Houston without a human ever touching the wheel. That trip was not a test. It was a paying commercial run operated by Aurora Innovation, which launched what it calls the first driverless trucking service on a U.S. interstate corridor. As of mid-2026, Aurora continues to expand those operations while the most prominent robotaxi program in the country remains shuttered and its parent company has moved on entirely.
The divergence is stark, and it comes down to a deceptively simple factor: roads. Highway trucking and city driving are two fundamentally different engineering problems, and the highway version turns out to be solvable years sooner.
Highway simplicity gives trucking a head start
Interstate highways are controlled-access roads. They have standardized lane markings, limited entry and exit points, consistent speed ranges, no pedestrian crosswalks, and no cyclists weaving between lanes. A truck running a fixed corridor between Dallas and Houston encounters a fraction of the unpredictable scenarios that a robotaxi faces on a single city block in San Francisco, where a jaywalker, a double-parked delivery van, and a cyclist making an unprotected left turn can all appear within seconds of each other.
Fewer of those “edge cases” means fewer chances for the software to fail. Fewer failures mean faster technical validation. And faster validation means an easier path through regulatory review.
Aurora built its entire commercial strategy around that insight. Rather than chasing the full complexity of urban driving, the company focused on hub-to-hub freight corridors in Texas, betting that highway-grade autonomy could reach commercial readiness while city-capable systems remained years away. Its SEC filings reflect that priority: capital allocation, milestone disclosures, and risk factors center on freight operations, not passenger vehicles.
Aurora is not alone in the corridor. Daimler Truck completed its acquisition of Kodiak Robotics to accelerate its own autonomous freight program. Torc Robotics, also backed by Daimler, has been testing on Southwestern interstate routes. Gatik operates autonomous middle-mile delivery trucks for retailers like Walmart. The competitive field is filling in fast, and the common thread is the same: highways first, cities later, maybe much later.
The timing aligns with an economic pressure point the trucking industry has struggled with for over a decade. Long-haul routes, which demand hours of monotonous highway driving far from home, are the hardest positions for carriers to fill and the likeliest to see turnover. The American Trucking Associations has repeatedly flagged driver retention as a top industry challenge. Autonomous trucks do not need rest breaks, are not subject to federal hours-of-service limits, and can theoretically run around the clock. That math has pushed major shippers and logistics firms to pour capital into the technology.
Robotaxis hit a wall that trucks drove around
While trucking programs quietly scaled up on interstates, the highest-profile robotaxi effort in the United States collapsed in public.
In October 2023, the California Department of Motor Vehicles immediately suspended Cruise LLC’s driverless testing and deployment permits after a series of safety failures, including an incident in which a Cruise vehicle dragged a pedestrian who had been struck by another car. The suspension pulled one of the most visible autonomous vehicle programs off the streets of San Francisco overnight. By December 2024, General Motors announced it would stop funding Cruise’s robotaxi operations altogether, writing down the investment and effectively ending the program.
The Cruise shutdown sent a signal that regulators and the public were not ready to tolerate safety lapses from cars carrying passengers through crowded neighborhoods. Waymo, operated by Alphabet, continues to run robotaxi service in San Francisco, Phoenix, and Los Angeles, but it remains the exception rather than the beginning of a wave. No other company has launched a comparable commercial robotaxi service in a major U.S. city, and proposed federal legislation to create a national framework for autonomous passenger vehicles has stalled repeatedly in Congress.
Autonomous trucking, by contrast, has not triggered the same level of regulatory alarm. The vehicles operate on road networks with fewer vulnerable users, at times of day when traffic density is lower, and on routes that can be selected specifically to minimize risk. That does not mean the technology is risk-free, but it does mean the failure modes are less likely to involve the kind of pedestrian-injury scenarios that ended Cruise.
The federal safety net covers both, but unevenly
The National Highway Traffic Safety Administration maintains a standing general order requiring crash reporting for all vehicles equipped with Automated Driving Systems or Level 2 Advanced Driver Assistance Systems. The order applies equally to autonomous trucks on I-35 and robotaxis in downtown Phoenix. Every operator must disclose incidents involving injury, fatality, airbag deployment, or certain other criteria within defined timeframes.
That shared reporting framework creates a data pipeline, but it has a significant limitation: NHTSA has not published a comparative breakdown separating trucking incidents from robotaxi incidents by route type, speed, or traffic density. The raw data exists in aggregate, yet without that granularity, it is difficult to confirm with hard numbers whether highway-based autonomous systems are producing meaningfully fewer reportable events per mile than their urban counterparts.
Academic researchers and insurance analysts have called for anonymized, route-level statistics showing incident rates per million miles, broken down by operating conditions. Without that detail, the policy conversation risks being shaped by the most dramatic failures rather than by a systematic picture of risk. The Cruise suspension dominates public memory of autonomous vehicle safety, even though it may not be representative of the technology’s overall track record.
What the data still cannot tell us
Aurora’s exact safety performance on Texas corridors remains opaque. The company has disclosed milestones through press releases and regulatory filings, but granular per-mile, per-route incident data has not appeared in public documents. Investors and regulators are working with broad corporate disclosures, not the kind of detailed statistics that would allow a rigorous comparison with urban robotaxi programs.
The same opacity applies across the industry. Daimler Truck, Torc, and Gatik have all announced progress without releasing the sort of independent, third-party safety audits that would let outsiders verify their claims. The companies argue, with some justification, that proprietary safety data is competitively sensitive. But the result is that the public is asked to trust corporate assurances at a moment when trust in autonomous vehicles is fragile.
Insurance and liability frameworks add another layer of unresolved questions. When a driverless truck is involved in a crash on an interstate, the question of who bears responsibility, the software developer, the truck manufacturer, the fleet operator, or the shipper, does not have a settled legal answer in most states. Texas, where Aurora operates, has been permissive toward autonomous vehicle testing, but permissive testing rules are not the same as a mature liability regime.
Two lanes, very different speeds
The picture that emerges by mid-2026 is not one of autonomous vehicles broadly arriving or broadly failing. It is a story of two lanes moving at very different speeds.
In one lane, autonomous trucks are hauling freight on controlled-access highways, generating revenue, and expanding their route networks with relatively little regulatory friction. The technical problem they solve is narrower, the economic incentive is immediate, and the road environment is forgiving enough to let imperfect software operate within acceptable safety margins.
In the other lane, robotaxis remain stuck behind a combination of technical complexity, regulatory caution, and public skepticism that the Cruise debacle intensified. Waymo’s continued operation shows the concept is viable, but scaling it beyond a handful of cities with favorable conditions has proven far harder than the industry projected five years ago.
The gap between the two may continue to widen. Every mile an autonomous truck logs on I-45 without incident builds the safety case for expanding to new corridors. Every month a robotaxi program stays grounded or limited erodes the political will to approve broader urban deployments. The technology is not inherently better suited to one application over the other in the long run, but right now, the highway is a far easier place to prove it works.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
