The reports began arriving in 2021, and by early 2022 the U.S. government’s auto-safety regulator had collected roughly 300 complaints about a troubling pattern: Tesla vehicles slowing down suddenly, for no apparent reason, while driving on highways. On Thursday, the National Highway Traffic Safety Administration closed the investigation, concluding that the problem no longer poses a meaningful risk to drivers. The case, which covered 695,000 Tesla Model 3 and Model Y vehicles, ended with a finding that the software fix Tesla released early in 2022 had worked.
The numbers support the regulator’s decision. Complaints about unintended deceleration fell from roughly 300 at the start of the probe to 45 in 2024, 19 in 2025, and just 3 so far in 2026, NHTSA said in its closure notice. The agency also found that the reported incidents did not change the vehicles’ lateral position in their lanes, and did not cause dangerous reductions in the distance to the car ahead. In regulatory terms, the risk of a crash never materialized, and the rate of reports has fallen so far that continued investigation was no longer justified.
The case had been a flashpoint in the broader debate over Tesla’s driver-assistance systems. The phantom-braking complaints, in which vehicles braked unexpectedly on open roads, were among the most frequently cited criticisms of Tesla’s camera-based Autopilot and Full Self-Driving systems, which rely on vision rather than the radar and lidar used by some rivals. The company has long argued that its software improves over time and that the complaints represented a small fraction of vehicles on the road. The closure of the probe, while not a formal vindication, supports that argument in the most concrete way available to a regulator.
The investigation had proceeded under an unusual shadow. NHTSA opened its preliminary evaluation in 2022, at a time when Tesla was already facing scrutiny over other safety questions, and the phantom-braking issue became a proxy for wider concerns about the company’s testing and release practices. The agency’s decision to close the case without a recall is a significant outcome for Tesla, which has resisted recalls it considered unnecessary and has at times clashed publicly with safety regulators.
The resolution also carries lessons for the industry. The case demonstrated how software can be both the problem and the fix: the defect was introduced by an over-the-air update and corrected by another update, with no physical part replacement required. That capability is a double-edged sword for regulators, who have struggled with how to apply recall rules written for hardware to vehicles whose behavior can change overnight. NHTSA has been developing guidance for software-driven recalls, and the Tesla case will be cited in that effort either way.
For Tesla, the closure removes a piece of regulatory overhang as the company prepares for a year of new products. The company delivered 480,126 vehicles in the second quarter, comfortably beating Wall Street expectations, and it has been expanding its lineup in the United States and Europe. The phantom-braking case was not the largest regulatory issue Tesla faces — the company is still responding to questions about its driver-assistance marketing and its crash data — but it was among the longest-running, and its conclusion is a small tailwind.
The timing of the closure also matters for how the industry reads NHTSA’s posture. The regulator has opened several investigations into Tesla over the years, ranging from autopilot crashes to steering failures, and the outcomes have been mixed. Closing the phantom-braking case with a finding that the risk was low sends a signal that NHTSA is willing to end investigations when the evidence warrants, a stance the broader auto industry has been pressing for.
Analysts noted that the incident-rate data was the decisive factor. The decline from hundreds of reports a year to single digits is exactly the pattern a regulator wants to see after a software fix, and the absence of crash consequences in the reported cases made the decision straightforward. The agency’s notice also credited Tesla’s update, released in early 2022, with addressing the underlying causes, which included conditions that could confuse the vehicles’ vision systems on highways.
The practical effect for Tesla owners is minimal — the fix was deployed years ago, and the vehicles covered by the probe continue to operate normally. But the case’s closure removes a data point that critics of the company’s software approach had cited repeatedly, and it provides Tesla with a regulatory reference it can point to in future discussions. The company has argued that its safety record, measured across millions of vehicles, is stronger than the headline complaints suggest, and a closed investigation is a form of documentation.
For the wider question of how self-driving software is regulated, the case ends where many industry executives hoped it would: with a finding that the software corrected its own problem, and that the regulator’s role was to verify, not to punish. The next chapter belongs to the companies deploying more advanced systems, and to the regulators deciding what evidence is enough to close a file.
