Two U.S. senators want federal regulators to pull apart the statistics Tesla uses to claim its driver-assistance software is far safer than a human behind the wheel. Senators Edward Markey of Massachusetts and Richard Blumenthal of Connecticut sent a letter on June 16, 2026 to National Highway Traffic Safety Administration (NHTSA) Administrator Jonathan Morrison, calling the analysis behind Tesla’s Full Self-Driving (FSD) numbers “weak and misleading” and an “urgent safety problem.”

The senators gave the agency until July 7 to answer a set of pointed questions, including whether NHTSA has ever independently checked Tesla’s claim that FSD is seven to ten times safer than human drivers, or requested the underlying crash data. They also want NHTSA to widen reporting rules for every company running automated driving systems, arguing the agency currently has no reliable way to confirm whether public safety claims match reality.

I have read enough of these NHTSA exchanges to know the questions matter more than the headline. One asks whether Tesla counts a crash only when FSD was engaged within five seconds of impact, against the thirty-second window in NHTSA’s own Standing General Order. The senators are asking, not asserting, but if that gap exists it decides which crashes appear in the data at all.

The Letter Targets Tesla’s Comparison Method, Not Just Its Numbers

Markey and Blumenthal argue Tesla inflates FSD’s safety record through the comparisons it chooses rather than through any single false figure. The letter flags three methods: measuring unlike crash outcomes against each other, stacking new Teslas against the much older average U.S. vehicle, and leaning on incomplete crash data.

The senators put it bluntly. Tesla’s claims “appear to rest on methodological choices that systematically inflate FSD’s apparent safety advantage,” they wrote, naming the company’s assertions that its vehicles are ten times safer and produce 85% fewer crashes than human-driven cars.

Two further questions go to data completeness. One asks whether Tesla’s reliance on automated telemetry drops crashes when connectivity fails or the car’s communication hardware is damaged in the collision itself. If the worst wrecks are the ones least likely to phone home, the dataset skews safe by design.

Reuters Reporting Triggered the Senate Request

The letter cites a Reuters examination from last month that found Tesla CEO Elon Musk and other executives increasingly citing figures they say prove FSD is up to ten times safer than people. Researchers interviewed for that report said Tesla compared the crash rate of FSD-piloted cars that triggered airbag deployments against a national crash rate covering far less severe accidents.

The age mismatch compounds the problem. Tesla measures its newer fleet against the average American vehicle, which is considerably older and lacks the automatic emergency braking and lane-keeping systems that have cut crash rates across the industry over the past decade. NHTSA confirmed it received the letter and is reviewing it. Tesla did not respond to requests for comment.

The Same Data Surfaced in Tesla’s European Approval Bid

The Senate scrutiny lands days after Tesla’s U.S. safety figures appeared inside its European regulatory submissions. Tesla handed Dutch and Swedish authorities self-published statistics, including a slide claiming FSD could have prevented 32,000 deaths and 1.9 million injuries, as EVXL reported on June 15. The Dutch road authority RDW, which approved FSD (Supervised) on April 10, told Reuters it does not rely on marketing claims and runs its own testing.

That distinction is doing heavy lifting. RDW granted its approval under UN Regulation 171 after more than a year and a half of independent testing, a process EVXL tracked from the first reported April 10 target date. The agency declined to say whether it assessed the U.S. numbers Tesla submitted alongside its own road data. Dutch transport minister Vincent Karremans defended the decision in parliament on June 16, telling lawmakers RDW relied on its own road testing rather than Tesla’s figures, as EVXL reported that day. Approvals already granted to the Netherlands, Lithuania, Estonia, Denmark and Belgium rest on RDW’s original type-approval.

NHTSA Already Has an Open FSD Investigation

The new request adds pressure to an agency already investigating the same software. NHTSA opened a probe into roughly 2.9 million Tesla vehicles equipped with FSD in October 2025, following 58 reports of traffic safety violations that included 14 crashes and 23 injuries. The agency later questioned Tesla’s “Mad Max” speed profile after drivers reported routine speeding.

Markey and Blumenthal have a long paper trail here, and their skepticism of Tesla predates this data dispute by years. They pushed NHTSA toward its September 2025 railroad-crossing investigation, applauded the 2023 recall of more than two million Autopilot-equipped Teslas, and called for an FTC probe of Tesla’s marketing in 2021. In January 2026, Markey introduced the AV Safety Data Act, which would force autonomous vehicle companies to report vehicle miles traveled, unplanned stoppages, and injuries.

EVXL’s Take

The fight here isn’t about whether FSD is good software. It’s about who gets to define the denominator. Tesla picks the comparison group, the crash threshold, and the reporting window, then publishes a ratio. That’s marketing dressed as science, and the disengagement-window question the senators raised is exactly the kind of framing choice that can do quiet work on a safety ratio.

I covered the European side of this story on June 15, and the throughline is consistent: when an independent body actually tests the car, like RDW did over eighteen months, it sets the marketing slides aside. The Dutch minister made exactly that argument in parliament the next day. The U.S. has never run that independent test on FSD’s safety statistics. That’s the real gap this letter exposes.

Here’s my call. NHTSA will respond by the July 7 deadline with an acknowledgment, but it will not commit to independently auditing Tesla’s safety math before the end of 2026. The agency has had the open FSD investigation since October 2025 and still hasn’t ruled. A letter from two senators in the minority won’t change that timeline. The data transparency fight gets won through the AV Safety Data Act or not at all, and that bill isn’t passing this Congress.

Sources: Reuters via GV Wire, Office of Senator Edward Markey statement.

EVXL uses automated tools to support research and source retrieval. All reporting and editorial perspectives are by Haye Kesteloo.