September 19, 2025

Tesla’s Optimus AI Leader Departs for Meta

Tesla’s robotics division is undergoing a leadership change as Ashish Kumar, who led artificial intelligence development for the Optimus humanoid robot, has left the company to join Meta as a research scientist. The move comes as Tesla places increasing emphasis on its AI and robotics programs, which CEO Elon Musk has said could eventually represent around 80% of the company’s future value.

Ashish Kumar’s Exit

Kumar announced his decision on X Thursday evening, describing his work on Optimus as “an incredible ride.” He highlighted the team’s progress in applying scalable methods, including reinforcement learning and training robots to improve dexterity by learning from video demonstrations.

“AI is the most significant bit to unlock humanoids,” he wrote.

Kumar also made clear that financial considerations were not a factor in his choice. “The financial upside at Tesla was significantly larger,” he posted. “If I wanted to optimize for money, I would have stayed at Tesla.”

His LinkedIn profile now lists him as a research scientist at Meta, where he will continue focusing on advanced AI development.

Leadership Transition at Tesla

The Optimus program remains a priority for Tesla, with oversight now falling to Ashok Elluswamy, Vice President of AI Software. Elluswamy has previously led development of Tesla’s Autopilot and Full Self-Driving systems, and his involvement reflects the company’s approach of creating a shared AI framework for both vehicles and humanoid robots.

Kumar’s exit follows the June departure of Milan Kovac, who had overseen the broader Optimus program. Despite these changes, Tesla continues to showcase new iterations of the robot, including “Optimus 2.5,” revealed earlier this month with improvements in motor control and walking speed. Musk has described the upcoming third-generation design as “sublime.”

Optimus in Tesla’s Broader Strategy

Tesla has positioned Optimus as a central pillar of its long-term growth, with plans to move from limited production in 2025 to potentially 100,000 units by 2026. Musk envisions the robots being deployed in a variety of settings, from manufacturing and logistics to agriculture and household tasks.

At the same time, Tesla has streamlined parts of its AI strategy, recently ending its in-house Dojo supercomputer program in favour of using chips from Nvidia and AMD, manufactured by Samsung. Musk said the move allows the company to focus on developing inference chips to support real-time decision-making in its AI systems.

Are you buying a Tesla? If you enjoy our content and we helped in your decision, use our referral link to get three months of Full Self-Driving (FSD).
Previous Article

Waymo and Via to Launch Autonomous Public Transit in Arizona

Next Article

Tesla robotaxi validation vehicles spotted in Phoenix

You might be interested in …