Elon Musk announced that Tesla has completed the design of its next-generation AI5 chip for Full Self-Driving capabilities. The chip will deliver performance comparable to NVIDIA's Hopper architecture, with two AI5 units matching a single Blackwell processor's power. This marks a significant advancement in Tesla's custom silicon development for autonomous driving.

The AI5 chip will be manufactured at both Samsung's Texas facility and TSMC's Arizona plant, representing a major win for Samsung's foundry business. Tesla's decision to split production between these facilities aims to diversify its supply chain and maintain control over chip availability for any demand scenario. The company is also sourcing LPDDR5X memory from SK hynix, with each AI5 chip featuring 192GB of memory across 12 modules.

Tesla has ambitious plans for rapid development cycles, targeting nine-month design periods for the AI6 chip. The company is also working on Dojo 3 and other custom silicon projects, suggesting an aggressive push into proprietary chip design. This strategy positions Tesla to reduce dependency on external suppliers while maintaining technological leadership in autonomous vehicle hardware.