Riders in San Francisco, Los Angeles, and Phoenix can now hail a robotaxi that will merge onto a freeway with no human behind the wheel. Waymo has expanded its fully driverless service to include freeways and interstates, a step that puts autonomous vehicles on roads where traffic moves at highway speeds and crash forces rise sharply. The expansion follows a series of state permitting milestones and adds new pressure on federal crash-reporting systems designed to track exactly these kinds of operations.
Why freeway-speed robotaxis change the safety calculus
Surface streets and freeways present fundamentally different risk profiles. At highway speeds, stopping distances grow, reaction windows shrink, and the energy involved in any collision multiplies. Waymo’s move onto these roads means its autonomous driving system must handle lane merges, high-speed lane changes, and mixed traffic that includes commercial trucks, all without a safety driver ready to intervene.
The California Department of Motor Vehicles maintains a record of Waymo’s approved areas, including driverless testing, deployment authorizations, and geographic expansions across the state. Those approvals now cover freeway segments, giving the company legal authority to operate at speeds that its earlier urban service areas did not include.
A reasonable expectation is that freeway operations will produce a higher rate of federally reported incidents per mile than surface-street driving during the initial months, driven less by software shortcomings than by the physics of speed differentials. When a robotaxi traveling at freeway pace encounters an unexpected lane intrusion or sudden slowdown, even a correct braking response may not prevent contact at energies far exceeding anything seen on city blocks. Whether the data bears this out will depend on how quickly federal reporting catches up to the expansion.
Permitting trail and federal reporting rules behind the rollout
Waymo’s freeway authority in California rests on a progression of state approvals. DMV records show the company moved through driverless testing permits, deployment permits, and successive geographic expansions under state autonomous vehicle rules. Each stage required the company to define an operational design domain, essentially a set of conditions and locations where the system is approved to function without a human driver. Freeways represent a new domain with higher speed limits and different traffic patterns than the urban grids Waymo has operated on for years.
On the federal side, NHTSA’s Standing General Order on crash reporting, formally designated SGO 2021-01, requires companies operating vehicles with automated driving systems to report crashes within set timeframes. The order covers both fully autonomous systems and Level 2 advanced driver-assistance systems. For Waymo’s freeway rides, every collision or incident meeting the order’s thresholds must be filed with the agency, creating a federal data trail that did not exist for these road types before the expansion.
The gap between what the data can show and what the public needs to know is real. SGO submissions are raw crash reports. They do not break out incidents by road type, speed at impact, or whether the vehicle was on a freeway versus a surface street. Without those variables, comparing freeway safety performance to urban performance requires analysis that neither NHTSA nor Waymo has published in a freeway-specific format. Riders and regulators are left to watch aggregate numbers and wait for patterns to emerge.
Where the expansion stands across three metro areas
The rollout is not uniform. Waymo expanded driverless service to freeways and interstates around San Francisco, where the company also gained curbside access at San Jose Mineta International Airport, according to the Associated Press. That airport pickup capability turns the robotaxi into a practical transit option for travelers, not just a novelty ride within city limits.
In Los Angeles, the highway rollout is gradual. The company has not opened all freeway segments simultaneously, instead phasing in coverage in a pattern consistent with its earlier city-by-city approach. Phoenix, where Waymo has operated driverless vehicles longer than in any other market, rounds out the three metro areas now seeing freeway service.
The geographic spread matters because each city presents different freeway conditions. San Francisco’s highways include steep grades, fog, and dense merge zones. Los Angeles freeways carry some of the heaviest traffic volumes in the country. Phoenix offers wide, flat highways but extreme heat that can affect tire performance and sensor calibration. Waymo’s system must handle all three environments without a human fallback.
Unresolved questions about freeway robotaxi safety
Several critical questions remain open. First, the California DMV’s approved areas page does not publish detailed maps or exact freeway segment coordinates tied to Waymo’s current permissions. Secondary descriptions confirm freeway authority exists, but the precise boundaries of where a driverless vehicle can and cannot operate on a given highway are not publicly accessible in map form. Riders must instead infer coverage from the app’s pickup and drop-off options, limiting independent scrutiny of how far freeway operations extend into complex interchanges or construction zones.
Second, neither state nor federal regulators currently require a freeway-specific safety performance dashboard. NHTSA’s standing order ensures that crashes are reported, but it does not compel companies to contextualize those incidents with exposure metrics such as miles driven on freeways versus surface streets. Without that denominator, the public cannot easily compare the risk of a robotaxi freeway trip to a human-driven one, or even to the same company’s city-only service.
Third, the transition zones between surface streets and freeways may pose unique challenges. On-ramps and off-ramps concentrate lane changes, speed adjustments, and driver expectations that can vary widely. Human drivers often rely on subtle cues, such as eye contact or informal zipper merges, to negotiate these spaces. How reliably an automated system interprets and responds to such behavior at scale is still an open empirical question, especially when other drivers are reacting to a vehicle they know has no human behind the wheel.
There are also questions about how quickly software updates can respond to emerging freeway-specific issues. If a pattern of near-misses appears in a particular merge configuration, regulators and the public have limited visibility into how that information is incorporated into new releases, or how thoroughly those changes are validated before being pushed to the fleet. The permitting framework assumes ongoing technical refinement, but does not spell out how freeway learnings should be communicated outside the company.
What regulators and riders should watch next
As freeway robotaxis become more common, several indicators will matter. Regulators can press for more granular reporting that distinguishes freeway crashes from urban ones, along with exposure data that allows for apples-to-apples comparisons. State agencies could also require clearer public maps of operational domains, so residents know where driverless vehicles are authorized to mix with high-speed traffic.
For riders, the near-term signals will be more experiential. Consistent lane positioning, predictable merging behavior, and transparent in-app explanations of routing choices will shape perceptions of safety as much as formal statistics. If freeway trips feel smoother and more conservative than the average human-driven ride, adoption is likely to grow even in the absence of detailed public data.
Waymo’s freeway expansion marks a turning point for autonomous vehicles, shifting the debate from whether robotaxis can handle city streets to whether they can safely share the fastest roads in the network. The permitting trail and federal reporting orders provide a framework, but not yet the freeway-specific insight that many stakeholders will want. Until that changes, the safety calculus for highway-speed autonomy will remain a work in progress, unfolding in real time on some of the most demanding stretches of asphalt in the country.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
