vLLM hits 500K GPUs as co-founder Simon Mo makes the case for open models

The UC Berkeley-born inference engine is now a cornerstone of production AI infrastructure, and its co-founder just raised $150M to prove itHalf a million GPUs are now running vLLM. It lets a GPU serve far more requests simultaneously by being smarter about how it allocates memory on the fly. Simon Mo, Inferact, and the $150M bet on open inferenceSimon Mo, one of vLLM’s core maintainers, co-founded Inferact in 2025. The real shape of 500K GPUsThe 500K GPU figure reflects the aggregate scale across all organizations running vLLM in production, spanning cloud providers, on-premise clusters, and enterprise deployments worldwide. Individual deployments typically operate across hundreds to thousands of GPUs, using a combination of tensor parallelism and pipeline parallelism.