Elon Musk
@elonmusk · May 28, 2026
SpaceX has almost finished writing V1.0 of an in-house AI training stack in C that exact-maps to 220k GB300s with 800G NICs, making heavy use of pipeline parallelism and getting as close to bare metal as possible.
The potential speed improvement vs JAX for large training runs is
Elon Musk
@elonmusk
Next will be writing the inference stack in C for simultaneous high-speed RL across a large block of GB300s.
(We do use a little C++ tbh, but not much)
04:40 PM · May 28, 2026 · 1.2M views
630
1.4K
8.7K