Inferact

Inferact commercializes vLLM, an open-source library for efficient large language model inference and serving, aiming to reduce costs and improve speed in AI inference. The technology leverages PagedAttention for memory optimization and higher throughput, supporting NVIDIA and AMD GPUs. Inferact launched in January 2026 with $150 million in seed funding at an $800 million valuation. The round was co-led by Andreessen Horowitz and Lightspeed Venture Partners, joined by Databricks Ventures and UC Berkeley Chancellor’s Fund. The underlying vLLM project was incubated in 2023 at a UC Berkeley lab of Databricks co-founder Ion Stoica, a member of Inferact's founding team. Inferact was founded in 2025 by Simon Mo, Woosuk Kwon, Kaichao You, and Roger Wang. Its business model centers on commercializes the open-source vLLM project through a VC-backed startup model.