FriendliAI
Software DevelopmentCalifornia, United States51-200 Employees
FriendliAI is a San Francisco-based software development company that positions itself as The Frontier AI Inference Cloud, offering an optimized engine for running frontier open-weight and custom AI models at production scale. The company emphasizes high GPU utilization to deliver performance advantages, claiming speeds up to three times faster than vLLM and cost savings of 50% to 90% versus closed-model APIs. Recent product and partnership activities include the March 13, 2026 launch of InferenceSense to monetize idle GPU capacity and the April 29, 2026 introduction of Day-0 support for NVIDIA's Nemotron 3 Nano Omni via dedicated endpoints for production-scale multimodal AI. In April 2026, FriendliAI announced integrations with Samsung SCP's GPU infrastructure to enable customers to run frontier open-weight models—such as GLM-5.1, MiniMax M2.5, NVIDIA Nemotron 3 Super, and DeepSeek v3.2—at maximum performance with production-grade reliability. The company has expanded its leadership and ecosystem, appointing Brian Yoo as Chief Business Officer in April 2026 and forming collaborations with Twelve Labs and NextDay AI to scale production inference and improve GPU efficiency.