RadixArk is a Palo Alto-based AI infrastructure startup that commercializes the open-source SGLang inference engine and the Miles reinforcement learning training framework, enabling large-scale, production-grade AI serving and post-training. Founded in 2025 by former xAI and NVIDIA engineers, it raised $100M in seed funding and supports deployments across hundreds of thousands of GPUs for customers including Google, Microsoft, and NVIDIA.
“Quite honorably, I can say that top engineering teams, in solving these engineering problems, often arrive at the same elegant solutions independently.”
Source→“Accel is pushing deeper into the stack alongside Nebius, including inference software from the ex-xAI team at RadixArk.”
AI-extracted from podcast / newsletter / paper summaries. May contain errors.