Tesla’s vision for autonomous transportation just got another crystal-clear reiteration straight from the company’s AI chief.
Speaking at the 2026 ScaledML Conference, Tesla VP of AI Ashok Elluswamy doubled down on what the company has been signaling for months now: the Cybercab is being built from the ground up for autonomy — and nothing else. As Elluswamy put it, “They are designed for autonomy. They don’t have any steering wheel, accelerator pedal, or brake pedal. It’s meant for full self-driving only. This will have the lowest cost of transportation, even beating public transport.”
That design philosophy is core to Tesla’s Robotaxi ambitions. The Cybercab is a two-seat, all-electric vehicle purpose-built for Tesla’s autonomous ride-hailing network, where vehicles can operate continuously without a human driver. Tesla owners will eventually be able to add their own cars to the fleet, but the Cybercab is meant to be the backbone — a clean-sheet vehicle optimized for cost, efficiency, and scale.
Importantly, while Tesla is already running validation Cybercabs on public roads and even highways, those early test vehicles still have steering wheels and traditional control pedals. Once volume production begins — currently expected around April — those controls are expected to disappear entirely, along with side mirrors, leaving the vehicle fully dependent on Tesla’s Full Self-Driving system.
Elluswamy also used the talk to reiterate Tesla’s long-standing, and often controversial, stance on vision-only autonomy. “It’s so obvious you can solve this with cameras. Why wouldn’t you solve with cameras? It’s 2026,” he said. “The self-driving problem is not a sensor problem, it’s an AI problem. The cameras have enough information already. It’s a problem of extracting the information, which is an AI problem.”
That philosophy underpins Tesla’s end-to-end neural network approach to autonomy — one that the company believes will scale faster and more efficiently than sensor-heavy alternatives. With Cybercab production trials already underway and real-world testing expanding across multiple U.S. states, Tesla appears to be moving quickly from theory to execution.
If you want a deeper technical dive, Elluswamy’s full ScaledML Conference talk is well worth a watch, especially his breakdown of Tesla’s foundational models for self-driving.
