Find ASIC Vendors

Tesla AI Chip Advancements

November 26, 2025

Get a Price Quote

Tesla has outlined further steps in its in-house silicon programme, with CEO Elon Musk saying the company is close to completing the tape-out of its AI5 chip and has already begun work on the subsequent AI6 generation. The update, shared on X on 23 November, suggests Tesla is continuing to expand its vertically integrated hardware platform for autonomous driving and robotics. For European engineers, the development is potentially relevant because Tesla’s silicon roadmap increasingly intersects with automotive compute, data-centre acceleration and next-generation manufacturing automation.

Design progress and production outlook

According to the announcement, the AI5 design has cleared its internal review and is positioned to deliver 2000–2500 TOPS, potentially offering around five times the compute capacity of the current AI4 generation. This level of performance is intended to support more advanced unsupervised autonomous-driving algorithms.

Musk stated: “Our goal is to bring a new AI chip design into mass production every 12 months. Ultimately, the number of chips we produce will exceed the combined output of all other AI chips,” highlighting Tesla’s ambitions for rapid iteration and volume scaling.

Engineering samples of the AI5 are expected in 2026, with volume production targeted for 2027. Tesla plans to deploy the chip across its vehicle platforms and data-centre systems, reflecting the company’s view that autonomous-driving models and robotics workloads will require sustained compute scaling.

Dual-foundry strategy and AI6 development

Tesla will maintain a dual-foundry approach. Now in early development, the AI6 chip is set to be manufactured in parallel at Samsung’s Taylor, Texas site and at TSMC’s facility in Arizona. Both the AI5 chip and AI6 chip use identical designs and software stacks, with physical differences arising only from process-level implementation at each foundry.

The company says these iterations are part of a broader strategy to support a growing hardware ecosystem covering autonomous driving, fleet data processing and robotics. This includes Tesla’s Optimus robot programme, where first-generation production is planned for late 2025 and a second-generation model is expected to follow in 2026.