Nimbus AI launches on-device multitask LLM for mobile and closes $45M Series A

AI · 5 min read

Nimbus AI today launched NimbusLite, a 300M-parameter multitask LLM designed to run locally on high-end phones while offloading heavier inference to a hybrid cloud. The company positions NimbusLite as a drop-in SDK for chat, summarization, code completions, and on-device vision-text fusion without sending user data to remote servers by default.

Alongside the product launch Nimbus announced a $45 million Series A led by Northbridge Capital with participation from Blue Harbor Ventures and Foundry Collective. The funding will accelerate model optimization across chipsets, expand the SDK to embedded platforms, and grow enterprise integrations into finance and health apps.

Early partners include a mid-market banking app that will pilot local document summarization and a health-tracking startup testing on-device symptom triage. Nimbus emphasized developer ergonomics — single-line integrations, privacy toggles, and transparent model cards — positioning itself against cloud-first incumbents in an increasingly regulation-conscious market.