AMD Ryzen AI NPUs Are Finally Useful Under Linux For Running LLMs

AMD Ryzen AI NPUs are now practically usable on Linux for running large language models, thanks to the release of Lemonade 10.0 and FastFlowLM 0.9.35. Lemonade 10.0 adds Linux NPU support for LLMs and Whisper, plus native Claude Code integration. FastFlowLM serves as the NPU-first runtime supporting context lengths up to 256k tokens on current-gen Ryzen AI hardware. Users need Linux kernel 7.0 or AMDXDNA driver backports, and the setup supports AMD Ryzen AI 300/400 series SoCs. This is a notable shift after years of the AMDXDNA kernel driver existing with minimal useful user-space software.