Frequently Asked Question

Why does almost all AI development happen on Linux?

Three reasons compound. The first is inheritance: training a large model is a high-performance computing problem, and the HPC community standardised on Linux decades ago. Every one of the world's 500 fastest supercomputers runs Linux, and the schedulers, parallel filesystems, and provisioning tools that clusters depend on were built for it.

The second is tooling. NVIDIA's CUDA platform, the drivers, the multi-GPU communication libraries, and the container integrations all matured on Linux first, because that is where the researchers were. GPU support is one of the few areas where Linux is not merely adequate but clearly ahead of the alternatives.

The third is culture. Machine learning research advances by reproduction, and a result that others cannot run is not useful. That makes open source a working requirement rather than an ideological preference, and it pulled the field towards the platform where open development was already normal.