Elon Musk told investors and followers this week that Americans will be using cars without human monitors on a wide scale before the end of 2026, per Reuters' write-up of his remarks. It is a forecast he has made in similar form for years. The question is whether anything in the regulatory or operational record has actually moved to support it this time.

What Musk said, and what is shipping

Musk's claim covers two distinct products. One is Tesla's consumer FSD (Supervised), which still requires a driver with hands available and eyes on the road. The other is the Robotaxi service that Tesla launched in Austin in June 2025 with a small fleet and a Tesla employee in the front passenger seat. Neither of those is a car without a human monitor. The Austin pilot has a monitor by design. The consumer product has a monitor by federal compliance.

For 'widespread' to be true in any normal reading, Tesla needs to remove the in-car safety rider in Austin, scale the geofence, and convince a regulator that the consumer stack can be run unsupervised. None of those three steps has a public timeline tied to a filing.

The federal file is open, not closed

NHTSA's Office of Defects Investigation opened preliminary evaluation PE24-031 in October 2024, covering roughly 2.4 million Teslas after four reported crashes in reduced-visibility conditions on FSD, including one pedestrian fatality. A preliminary evaluation is the first step. It can escalate to an engineering analysis and then to a recall. It has not been closed. Until it is, Tesla cannot credibly argue to federal regulators that the same software is ready to drive without supervision.

Separately, the California DMV's case alleging that Tesla misled customers on Autopilot and FSD capabilities remains live. California is the largest US autonomy market. A loss there constrains any unsupervised rollout in the state where Tesla is headquartered.

What 'widespread' looks like when someone else is doing it

Waymo is the only US operator running passenger service with no human in the vehicle at scale. Its public safety hub reports cumulative rider-only mileage and per-mile crash comparisons against human-driver baselines in the cities it serves. Waymo's path from a driverless rider pilot to commercial operations across several cities took years of incremental scaling, not months. That is the pace of the only company that has actually done what Musk is promising. It is not a pace consistent with going from a supervised Austin pilot to nationwide unsupervised use inside seven months.

AutonomyEV's opinion

Take the prediction at face value and the math does not work. Tesla would need to clear PE24-031, satisfy California, remove the Austin safety monitor, publish per-mile incident data of the kind Waymo discloses, and scale a fleet, all in months. Tesla has done none of those things publicly. The more interesting read is operational: the Austin pilot is the data-gathering exercise Waymo ran years before going driverless. Treat 2026 as the year Tesla learns what unsupervised service costs to run, not the year it ships one to everyone. If you are an investor, a city transport official, or a fleet buyer planning around Musk's timeline, plan around the filings instead.