Overview

AI-first autonomy, built on Arm

In the era of physical AI, vehicles are no longer just machines — they are intelligent agents capable of real-time perception, reasoning, and action. Tensor has built the world’s first personal RoboCar and AI-agentic vehicle around this principle.


Powered by Arm-based compute platforms, the Tensor RoboCar processes millions of data points per second, delivering a Level 4-capable vehicle that can safely navigate complex environments while enabling natural, language-based human interaction inside the cabin.


Together, Tensor and Arm are proving that scalable, energy-efficient compute is the foundation of the autonomous future.

Learn about Automotive on Arm

Impact

Result

  • 8,000 TOPS centralized autonomy compute
  • 25.6 million lidar beams per second processed
  • 433 Arm-based cores per vehicle

Market / Use Case

Powered by Arm

“Delivering personal autonomous vehicles at scale requires more than breakthrough AI and autonomy, it demands advanced engineering for safety, redundancy, reliability, and power efficiency,”
Dr. Jewel Li, Tensor COO
Autonomous vehicle driving past modern urban buildings.
Challenge

Building a RoboCar around physical AI

Autonomous driving demands more than raw compute—it requires precision. From the moment a photon hits a camera sensor to the torque applied to a steering actuator, every millisecond matters.

AI tensor supercomputer board delivering 8,000 TOPS performance.
Solution

Tensor designed its RoboCar around this end-to-end latency challenge. With 17-megapixel cameras and five lidars, including a top-mounted lidar emitting 25.6 million beams per second, the system processes massive volumes of data in real time. At its core is an 8,000 TOPS automotive-grade supercomputer built on Arm-based CPUs, spanning edge sensors to centralized AI systems.

Rather than adapting AI to a vehicle, Tensor shaped the vehicle around AI. Sensor placement defined the design. Compute requirements defined the architecture. The scalable Arm CPU platform unified it all on a single architecture, delivering a cohesive AI-first platform that blends performance, efficiency, and safety.

Key Takeaways

  • Tensor built the world’s first AI-agentic personal RoboCar powered by Arm-based compute across the entire vehicle.

  • 8,000 TOPS of automotive-grade compute enables real-time Level 4 autonomy.

  • Arm architecture unifies sensors, centralized compute, and agentic AI systems.

  • Physical AI demands ultra low latency from perception to actuation, enabled by the scalable Arm platform.

  • Autonomous vehicles built on Arm can reduce time, improve safety, and enable private AI ownership.

Related Stories

Explore similar stories

Google automotive AI powering automotive transformation

See how Google Cloud and Arm are bringing generative AI to the entire automotive value chain, from manufacturing to in-vehicle experiences.

Learn more

Lenovo accelerating L4 autonomy with Arm-based computing

Discover how Lenovo’s L4 domain controller, built on Arm-based NVIDIA DRIVE AGX Thore, delivers over 2,000 TOPS of AI compute for scalable autonomous mobility.

Learn more

Geely creates next-generation driving experiences built on Arm

Explore how Geely’s Arm-powered intelligent cockpit delivers real-time AI, natural language interaction, and cross-domain integration accross cockpit and driver-assistance systems.

Learn more
Discover more success stories