Tesla sets 400 MW AI compute target by year-end, confirms Semi and Optimus production

Tesla says Semi production has begun in Nevada and Optimus production in Fremont, with a 400 MW AI training compute target by year-end.

Tesla said it expects rising data center demand to keep driving demand for its products. The company said production has begun for the Tesla Semi in Nevada and for Optimus in Fremont, and that it plans to reach 400 MW of AI training compute by year-end to support development of both its vehicle and robotics programs.

The update addresses part of an execution concern that has surrounded Tesla's Optimus program since September, when Electrek, citing reporting from The Information, said the robot's hand and forearm contain more than 100 screws and small components still assembled by hand, and that fixtures struggle to line up parts built to tighter tolerances than those used in cars, causing rework. The same reporting said Tesla relies on outside suppliers, many based in China, for the motors and precision gears that drive the joints, and that some of those suppliers can produce good parts in prototype quantities but struggle to hold quality steady at higher volume. It also said some of the hand's touch sensors have had reliability issues, which Tesla plans to address next year with a replaceable 'sensing glove' rather than swapping out the entire hand.

Confirming Semi production in Nevada and Optimus production in Fremont speaks to part of that execution concern, though it does not resolve the hand and supplier issues flagged in that reporting. The 400 MW compute target ties Tesla's AI infrastructure spending to both its vehicle and robotics efforts, treating data center capacity as a bottleneck shared across both rather than one specific to cars.