Tesla's Cybercab reveals a dangerous gap between ambition and safety

The Probe Already Started

The National Highway Traffic Safety Administration has opened an investigation into Tesla’s Cybercab rollout. That news arrived alongside a product reveal that barely felt like a product reveal — a quiet, invite-only gathering where Elon Musk apparently didn’t even give a keynote, and the official presentation lasted roughly fifteen minutes.

The gap between the two moments tells the story. Tesla is building a vehicle it claims will transform transportation, then handing it over to regulators mid-build and hoping they adapt to its timeline rather than the other way around.

The Cybercab is supposed to be a fully autonomous ride-hailing car that runs on cameras and AI, with no steering wheel and no human driver. Tesla says it will cost far less than a Waymo vehicle to operate. The company is also insisting it doesn’t need the same safety architecture as conventional cars — a position that has already attracted federal scrutiny.

What the Cybercab revealed in its sparse documentation is a pattern of decisions that look smart on a cost sheet and dangerous in practice.

No Kids, No LATCH, No Explanation

One of the most quietly alarming details in the Cybercab paperwork is that children under 13 are not allowed to ride in the vehicle “at this time.” The phrase matters. It suggests the restriction is temporary, not permanent — but also that Tesla has not figured out how to make the car safe for young passengers.

Older children, ages 8 to 17, can ride in Tesla’s Robotaxi Model Y SUVs, which still have child seat anchors. The Cybercab does not. It lacks the standard LATCH system entirely. Child seats can only be secured with the seat belt, and even then, the car’s interior geometry and door design make that a awkward proposition.

All riders under 18 must be accompanied by an adult in both vehicles, according to Tesla’s terms. So the practical effect of the Cybercab ban is that families with young children simply cannot use the service — at least not for now.

This is a striking choice for a company whose CEO has publicly discussed wanting more children. It is also a signal of how lightly Tesla is treating a basic question of public transit: who is this vehicle designed for, and who gets left out?

The Door Problem Wasn’t New

Perhaps the most consequential detail concerns the doors. The Cybercab uses electronic latches — no traditional handle, no mechanical fallback built into the primary opening system. When a ride begins or ends, the doors open automatically. There is also a small exterior button that riders can press to unlock them.

Inside, Tesla has placed the manual emergency release on the armrest of each door. The company says it is easy to find. It may be. But the broader issue remains unresolved.

Tesla is already facing a massive recall in China — 3 million vehicles pulled from the road last month — because of exactly this problem. An NHTSA-style investigation there found that electronic door latches could trap occupants after a crash. The recall involved Tesla and a number of other automakers. The fix has not been applied across Tesla’s fleet.

Now the company is putting the same unproven door system into a vehicle that will not have a driver inside to open them in an emergency. If the electronics fail after a collision — and they can — the passengers are locked in a metal box with no one inside the car to help them.

Cost-Cutting as Policy

The Cybercab’s brake system makes the same calculation. Tesla is using brake-by-wire, an electronic system that replaces hydraulic brake lines with electric actuators at each caliper. Musk framed it as a simplification, writing that it avoids routing plumbing around the car.

It also avoids a layer of mechanical redundancy. In a conventional braking system, if the electronics fail, a physical hydraulic connection still lets the driver stop the car. In the Cybercab, there is no driver. If the brake-by-wire system fails, there is no mechanical backup — only software.

The same logic already appeared in the Cybertruck, which was the first mass-market vehicle to use steer-by-wire, severing the physical link between the steering wheel and the front wheels. Tesla is treating electronic redundancy as an optional upgrade rather than a baseline requirement.

The rationale is cost. The Cybercab was designed from the ground up to be cheap — cheaper to build, cheaper to operate, cheaper to scale. Every component that can be simplified is simplified. The assumption, implicit but clear, is that the software will compensate for whatever hardware complexity is removed.

Windows That Won’t Open

Then there are the windows. They cannot be fully opened. Tesla’s documentation does not explain why.

In a car with no driver, no door handles you can pull manually, and a brake system that lives entirely in software, the inability to open a window is a small detail that carries disproportionate weight. It removes one more escape route. It also raises a question Tesla has not answered: what happens when the software controlling the cabin environment fails and the passengers need fresh air, or a way to signal for help?

The USB-C ports, by contrast, seem almost generous — 90 watts, four times what most cars provide. But that is a convenience feature in a vehicle where the fundamental safety architecture is still unproven.

What Happens Next

The NHTSA investigation is the first real test. Federal regulators have broad authority to demand recalls, change certifications, and block deployments if they conclude a vehicle design poses unreasonable safety risks. The Cybercab’s current configuration — no driver, electronic doors, brake-by-wire, no child restraints — gives regulators plenty to scrutinize.

Tesla’s strategy has always been to ship first and iterate later. The company applied that playbook to software updates and it worked, mostly. It is less clear that the same approach works when the product is a two-ton metal machine moving through public streets at 40 miles per hour.

Other cities are watching. Waymo operates in a handful of American cities under strict regulatory oversight. Cruise was shut down after a pedestrian dragging incident. Zipcar, Uber, and Lyft are all studying autonomous taxi fleets but have not committed to their own hardware. The Cybercab is the first fully driverless vehicle to reach this stage of public rollout without a manual fallback inside the car.

If Tesla succeeds, it changes the economics of urban transit. If it fails — and the design choices suggest failure is a realistic possibility — the regulatory backlash could slow autonomous vehicle adoption everywhere, not just in California or Texas.

The no-kids policy is the easiest detail to dismiss as an edge case. But it is also the clearest evidence that Tesla has not thought through who rides in its cars and what happens when things go wrong. The doors, the brakes, the windows — each one is a decision made in favor of cost over redundancy. None of them are fatal on their own. Together, they form a vehicle that is aggressively optimized for one thing: getting to market faster than safety can catch up.

That strategy may work for a while. It will not work forever.