S1677 would require New Jersey robotaxis to combine cameras with at least two other sensing technologies, a rule that quietly excludes Tesla's camera only stack and could spread to other states.
New Jersey's S1677 doesn't name Tesla. It names sensors.
The bill, advancing in the 2026 session, would require any fully autonomous vehicle on state roads to combine cameras with at least two other sensing technologies, in most readings lidar and radar. That hardware floor is what gives the bill reach beyond a single company.
S1677 is sponsored by Sen. Andrew Zwicker (D-NJ). His framing is careful. "This is not anti-Tesla. I'm pro-New Jersey safety," he told The Verge. Asked whether camera-plus-software alone could meet the bar, Zwicker was direct: "Can we get there? Maybe. But we're not there yet." The bill text, available as a 2026-session PDF from the New Jersey Legislature, is the architectural move: define safety in hardware terms and let the AV industry sort out which stacks qualify.
Carnegie Mellon engineering professor Philip Koopman, who has spent two decades on autonomous-vehicle safety, lays out the technical case. "To run 24/7 across the majority of public roads in New Jersey today, it needs lidar," he said. "It's pretty clear that today camera-only technology is not up to the challenge." Lidar, which maps a scene with laser pulses, is what most non-Tesla robotaxi fleets use alongside cameras and radar. Waymo and Cruise both run multi-sensor stacks. Tesla removed radar from its production cars in 2021 and has never deployed a lidar sensor.
Musk's counterargument is the one that could eventually falsify the mandate. His bet is that a vision-only neural network trained on enough miles will outperform any sensor fusion stack, and that the accident record will vindicate the choice. He has framed the radar removal as a safety move, citing what he called "sensor contention" between cameras and radar in a post last year. If vision-only matches or beats multi-sensor fusion in the next few model cycles, S1677's hardware floor will look in hindsight like a state legislature freezing a winning architecture out of its market. That is the test the bill sets up, and the reason the dispute is real rather than theatrical.
The fleet counts behind that bet are lopsided. Tesla is operating a small number of fully autonomous Robotaxis on public Texas roads, per Electrek's coverage of S1677. Waymo operates more than 3,500 driverless vehicles across 11 U.S. metro areas, according to the same reporting. Musk has predicted Tesla's fleet will grow to "hundreds of thousands by the end of this year," a figure that has not been independently verified. The gap between the current Tesla footprint and Musk's year-end promise is the bet S1677 is testing, and the next state to file a sensor-mandate bill will tell you which way the regulatory geography is tilting.
New York is weighing a nearly identical hardware mandate, according to The Verge's reporting on S1677. New Jersey has a long legislative paper trail on autonomous vehicles, including the S4702 bill text tracked by LegiScan from the prior session, which shows the state refining AV rules across multiple cycles. A sensor-diversity mandate is a portable tool for state lawmakers: it constrains a specific architecture without ever naming a company, so it survives lobbying and electoral pressure in a way a "ban Tesla" bill would not. That portability is what makes S1677 worth tracking outside New Jersey.
The watch item is the bill itself. S1677 still has to clear committee, a floor vote, and the governor's desk before any sensor floor takes effect, and the current 2026-session PDF status as of late July 2026 is the next concrete thing to track. If it passes, expect two filings within a quarter: a Tesla lawsuit arguing the mandate is an undue burden on interstate commerce, and a second state legislature introducing the same language under a different bill number.