The expansion follows early implementations by companies like Legato and Marshall, which utilized the stack to integrate advanced audio processing into compact hardware. The femtoAI ecosystem includes the Sparsity Studio measurement tool, proprietary compilers, and support for both custom models and pre-built architectures like OpenAI Whisper. By moving beyond a closed beta, the company aims to accelerate the deployment of AI in edge devices ranging from smart glasses to household appliances.
In section Releases
femtoAI Opens Full-Stack Silicon and Compression Platform to Developers
After a successful beta phase, San Bruno-based startup femtoAI is opening its developer community to the public. The platform provides engineers with direct access to patented neural-network sparsity tools, including the Sparse Processing Unit evaluation kit, designed to slash AI power consumption and memory footprints by up to 90 percent.
Chief Business Officer Bill Hoppin noted that the platform has already provided a competitive edge for early enterprise partners. The company, founded in 2018 by researchers from Stanford’s Brains in Silicon Lab, emphasizes that these efficiency gains do not compromise model accuracy. Developers can now access the full suite of tools and documentation, with the company hosting a community event during SF Tech Week to further engage the engineering ecosystem.
Comments (0)
No comments yet. Be the first!