vLLM maintainers founded Inferact and raised $150 million in January 2026
Investor a16z says the open-source AI inference engine runs on about 500,000 GPUs at any given time.
TLDR
a16z says four vLLM maintainers founded Inferact in January 2026, raising a $150 million seed round at an $800 million valuation. The investor traces vLLM to a 2022 Berkeley project that used virtual-memory techniques to manage GPU requests. It says the software now has more than 3,000 contributors and runs on roughly 500,000 GPUs at a time. Inferact is hiring as it develops software for serving AI models.
Combined views
—
1 Source, first seen ago
