Tesla Shifts AI Focus to Enhance Inference Chip Development

Tesla, led by CEO Elon Musk, is redirecting its artificial intelligence research to prioritize the development of inference chips, crucial for enabling real-time decision-making in AI systems. This announcement follows reports that Musk has ordered the closure of Tesla's in-house Dojo supercomputer team, as revealed by Bloomberg News.

The Dojo supercomputer, initially crafted to train Tesla's autonomous-driving system through custom training chips, is no longer the company's focus. Musk explained in an online post that maintaining two separate AI chip designs is not viable. The company instead will concentrate its resources on the AI5 and AI6 chips, aiming to excel in inference while remaining competent in training applications.

This strategic pivot comes amid Tesla's comprehensive restructuring, which includes significant executive turnover and job cuts. Musk is realigning Tesla's objectives toward AI-centered self-driving technology and robotics, with future AI6 chips planned for use in autonomous vehicles and Optimus humanoid robots. Notably, Tesla has secured a $16.5 billion arrangement with Samsung Electronics for chip sourcing, indicating substantial long-term investments in its AI roadmap.

(With inputs from agencies.)