California has set new operational rules for robotaxi companies, putting a clear emergency-services obligation alongside the technology’s broader public-road ambitions. Governor Gavin Newsom has signed Senate Bill 1246, a measure that allows local penalties when an autonomous vehicle blocks first responders for more than 30 minutes during an emergency.

The law is not immediate. Its requirements take effect on July 1, 2028, giving operators time to change procedures, staffing and communications systems. But its purpose is direct: a driverless vehicle cannot become an additional obstacle when ambulances, fire trucks, police or other responders need a clear route.

That sounds like a basic expectation, but it addresses a difficult practical question for autonomous vehicles. A conventional car that has stopped in the wrong place generally has a driver who can move it, take directions from an officer or explain what has gone wrong. A robotaxi must instead recognize the situation, safely execute an appropriate response, or quickly connect the people at the scene with someone capable of helping. SB 1246 is built around making those fallback paths more concrete.

What SB 1246 requires

The emergency-blockage provision is the most immediately understandable part of the legislation. Local authorities will be able to penalize an operator when one of its robotaxis blocks first responders for longer than 30 minutes in an emergency. The threshold matters because it turns a broad public-safety concern into an enforceable standard aimed at prolonged obstruction, rather than treating every unusual stop as automatically punishable.

For more context, explore Thunderbolt vs. USB-C on Macs: Same Connector, Very Different Capabilities.

The bill also requires operators to provide on-site emergency contacts known as local incident technicians. These workers are intended to assist immediately when autonomous vehicles are involved in accidents or obstruct roads. In plain terms, the rule calls for a local human layer between an automated fleet and the emergency personnel dealing with an incident.

That is distinct from simply having a customer-service number or a remote operations center. A local incident technician is an on-site contact: someone emergency responders can call upon when an autonomous vehicle needs assistance in the physical world. The supplied details do not specify response-time targets, staffing levels or how companies will organize those technicians. Still, the policy direction is clear. Companies will need a practical answer to the question of who comes to the scene when a vehicle cannot promptly resolve a problem by itself.

Operators will also be required to use remote human drivers based in the United States who hold US driver’s licenses. This requirement is notable because remote support is often discussed as though it means an off-site person continuously pilots every robotaxi. That is not necessarily how these systems operate.

Remote assistance is not the same as continuous remote driving

“Remote human driver” can describe a person able to take control or provide support from a distance, but SB 1246’s relevance is most apparent in exceptional situations: obstructions, accidents, complex road conditions and system problems. The rule does not mean that a robotaxi is ordinarily being manually driven from afar throughout each trip.

Tesla’s described approach is to allow remote human control in difficult situations only after available intervention actions have been exhausted. Waymo has said its vehicles can at times rely on guidance from remote human operators when they are stuck; those operators have been located in other countries. California’s new requirement adds a geographic and licensing condition for the remote human drivers covered by the law.

That distinction is important for judging what the bill changes. It is not a command to replace automated driving with a hidden remote chauffeur. It is a requirement focused on the people available when automation needs help. If an AV becomes immobilized or contributes to an obstruction, the quality of the operator’s human escalation process can be as consequential as the vehicle’s software response.

The legislation also imposes information-sharing duties during system-wide failures. Companies must notify local authorities about the location and status of their vehicles when such failures occur. A system-wide failure means a problem affecting a fleet or service system rather than a single isolated car. For emergency planners, knowing where affected vehicles are and what state they are in can be crucial: a stopped vehicle, a vehicle awaiting help and a vehicle attempting to clear an area create different operational issues.

Finally, operators must provide additional reporting on how quickly they respond to incidents involving their vehicles. Reporting alone does not remove a blocked lane, but it creates a record that can show whether a response system is functioning as intended. Over time, the data could help local authorities distinguish between an occasional difficult edge case and a recurring failure to support emergency operations.

Why emergency interference has become a central concern

The law follows several high-profile situations that illustrated how citywide conditions can create problems for autonomous fleets. In December 2025, multiple Waymo vehicles became stuck in San Francisco streets after a power outage disabled traffic lights. When signal infrastructure disappears, the road environment changes abruptly. Vehicles that are designed to make decisions using normal traffic-control cues can encounter an unusual condition across a broad area at once.

Another San Francisco incident occurred during Fourth of July celebrations this year, when heavy traffic left Waymo vehicles stranded after they ran out of power. The vehicles had to be towed. These events are not identical: one involved a power outage affecting traffic lights, while the other involved congestion and vehicles losing power. Yet both point to an operational reality beyond normal lane-following: fleets must cope with disruptions that can leave multiple cars unable to proceed.

Outside California, a Waymo vehicle in Austin, Texas, blocked an ambulance responding to a shooting. The US National Highway Traffic Safety Administration had already asked autonomous-vehicle operators to address what it identified as a clear pattern of driverless AVs interfering with law enforcement and other first responders.

The concern is not that every robotaxi encounter will create a 30-minute blockage. Rather, emergency services cannot plan on that assumption. Their work depends on reliable access, fast decisions and predictable coordination during events that may already be chaotic. A stalled automated vehicle may be manageable in ordinary traffic; in an emergency route, the same problem can have far greater consequences.

A law about operational accountability

SB 1246 moves the debate away from broad claims about whether autonomous vehicles are safe in general and toward specific operational responsibilities. It asks operators to account for the moments when vehicles are confused, stuck, disabled, depleted of power or affected by a larger system problem.

There are several connected layers:

  • Physical access: vehicles must not leave first responders obstructed for an extended period.
  • Local intervention: emergency personnel must have access to an on-site contact through local incident technicians.
  • Remote support: covered remote human drivers must be US-based and hold US licenses.
  • Fleet visibility: authorities must receive vehicle location and status information during system-wide failures.
  • Measurable performance: companies must report response times for incidents involving their vehicles.

Each layer addresses a different kind of uncertainty. A penalty creates a consequence for a prolonged blockage. A local technician addresses the need for physical help. Remote-driver rules address who may perform distance-based intervention. Failure notifications address coordination when a problem affects many vehicles. Response-time reports address whether the operator’s stated procedures work promptly in practice.

For communities hosting robotaxi operations, that package could make interactions with fleets more structured. For companies, it means the service cannot be defined solely by the autonomous driving system inside each vehicle. The surrounding human operation—incident staffing, contact channels, remote support and failure communication—is part of the service as well.

What the July 2028 start date means

Because the law does not take effect until July 1, 2028, operators have a substantial compliance window. That interval gives them time to establish local incident-technician arrangements, ensure qualifying remote-driver practices, develop notification procedures and prepare the incident-response reporting the measure requires.

It also means California residents and emergency agencies should not treat the signing as an instant change to today’s robotaxi response practices. The rules are enacted, but their enforceable start is later. The lead time appears designed to make compliance feasible across an operational model that combines vehicles, remote teams, local staff and public agencies.

The larger policy message is already apparent. Autonomous fleets are being judged not only on what happens during an ordinary ride, but on what happens when routine conditions break down. Traffic-signal outages, packed streets, emergency scenes and immobilized vehicles are all situations where a robotaxi operator’s support network can matter as much as the vehicle’s automated decisions.

California’s SB 1246 does not eliminate those difficult situations. It establishes expectations for how operators must prepare for them—and gives local authorities a route to act when a robotaxi leaves first responders waiting too long.