Stability AI and Arm Bring On-Device Generative Audio to Smartphones

Stability AI partnered with Arm to run Stable Audio Open directly on smartphone CPUs, enabling offline text-to-audio generation.

The optimized model is about 30 times faster, cutting an 11-second clip’s generation time from 240 seconds to under 8 seconds.

The system uses Arm software tools including KleidiAI, ExecuTorch, and XNNPack, and will be shown at MWC Barcelona.

The move highlights how generative AI can work efficiently on consumer phones without cloud access, lowering latency and broadening creative audio tools.