ScaleGrid raises $55M to commercialize federated inference for edge AI

AI · 6 min read

ScaleGrid announced a $55 million Series B led by InfraGrowth Capital and released a federated inference fabric designed to orchestrate model execution across heterogeneous edge fleets and cloud endpoints. The fabric focuses on privacy-preserving workloads and reduces bandwidth by executing as much inference as possible locally while sharing only compact summaries.

Customers in smart retail and connected vehicles can push updated submodels regionally without shipping full models, enabling rapid adaptation to local conditions. ScaleGrid also offers tools to profile device capabilities and auto-split models for optimal placement.

Use of funds includes broadening hardware support, adding developer tools for debugging distributed inferences, and building partnerships with chipmakers for optimized runtimes. Investors highlighted the commercial potential of consistent low-latency inference across millions of devices.