Ex-Tesla Optimus AI chief launches startup focused on non-humanoid industrial robots
Ashish Kumar, who led AI for Tesla's Optimus humanoid program, has co-founded Intelligent Machines. The San Francisco startup is developing task-specific industrial robots for manufacturing and aerospace, using reinforcement learning. It is raising a seed round that could reach about $100 million and has hired several former members of Tesla's Optimus team.
Kumar’s move follows about a year leading Optimus AI and a prior short stint at Meta. His new venture, Intelligent Machines, targets narrow industrial and aerospace duties rather than a human-shaped generalist, and it has already added multiple ex-Optimus colleagues.
The startup’s approach relies on reinforcement learning, a trial-and-error method Kumar studied at UC Berkeley. Its planned seed round could be unusually large for an early-stage company. Meanwhile, Tesla’s Optimus effort has faced slow task learning despite extensive training data, and another senior leader left in mid-2025.
If specialized robots prove reliable, they could change demand for skilled manufacturing and aerospace labor, potentially automating some narrow tasks while creating roles in robot deployment, maintenance, and training. Smaller firms may gain access to automation suited to narrow tasks, though limited per-design markets could keep costs high. Tesla’s humanoid effort may face more competition for talent. The outcome depends on whether reinforcement learning transfers efficiently and reliably from simulation to factory floors.