PrismSense launches on-device multimodal SDK and secures $20M bridge round for edge AI

AI · 4 min read

PrismSense's SDK bundles optimized vision-language models, audio event detectors, and sensor fusion primitives packaged to run on ARM NPUs and dedicated vision accelerators. The kit includes model quantization pipelines and privacy-preserving inference layers.

The $20 million bridge round was led by Aperture Capital with follow-on investors from several hardware partners. PrismSense says the funds will accelerate certification on additional silicon platforms and expand their device-reference library.

The SDK emphasizes offline operation and encrypted model containers so enterprise customers can deploy sensitive models without exposing data to the cloud. PrismSense also introduced an evaluation suite for measuring latency, power, and throughput across devices.

Early customers are pilots in industrial inspection and retail analytics. PrismSense plans a developer program and expanded tooling for model explainability on-device to support regulatory requirements in safety-critical applications.