Nota AI Signs AI Optimization Technology Supply and Strategic Partnership Agreement with Mobilint Accelerating Commercialization of On-Device AI on Domestic NPU Infrastructure

Why It Matters

The collaboration creates a domestic, end‑to‑end AI stack that accelerates commercialization of on‑device inference, reducing reliance on foreign semiconductor ecosystems and opening new revenue streams in high‑growth edge markets.

Key Takeaways

• Nota AI supplies NetsPresso® to Mobilint’s MLA100/MLA400 NPU cards.
• Partnership targets edge video AI markets like smart city and industrial safety.
• NetsPresso® optimizes AI models for low‑power, on‑device inference.
• Collaboration expands Nota AI’s footprint from mobile/data‑center to edge deployments.
• Joint marketing aims to commercialize Nota Vision Agent on domestic NPU hardware.

Pulse Analysis

The rapid rise of edge computing has pushed AI inference closer to sensors, demanding software that can squeeze models into limited power and memory budgets. Nota AI, a Korean firm specializing in model compression, announced a strategic supply agreement with Mobilint, a domestic NPU designer. By licensing its NetsPresso® platform for the MLA100 and MLA400 accelerators, Nota AI brings proven optimization tools to a hardware line aimed at real‑time video analytics. The deal signals a maturing domestic AI stack that can compete with imported solutions from Arm and Nvidia. The agreement also includes joint R&D funding to accelerate firmware integration.

NetsPresso® automates quantization, pruning and layer‑wise tuning, delivering up to 3‑times faster inference on Mobilint’s low‑power chips. For edge video workloads—industrial safety monitoring, intelligent transportation, smart‑city surveillance—the combined package offers a single‑vendor path from model training to deployment. Mobilint’s MLA series, built on a 7 nm process, delivers high throughput with a modest thermal envelope, while Nota’s software ensures that complex vision models such as the Nota Vision Agent (NVA) run reliably on‑device. Early benchmarks suggest a 30‑40 % reduction in latency compared with unoptimized models. Customers can access a cloud‑based profiling tool that visualizes performance gains across layers.

The partnership expands Nota AI’s ecosystem beyond mobile and data‑center customers to the edge, completing a three‑tier coverage strategy. As Korean manufacturers seek to reduce reliance on foreign IP, the collaboration provides a home‑grown alternative for sectors that value data sovereignty and low latency. Analysts expect the joint offering to accelerate commercialization of on‑device AI in Korea’s smart‑factory and smart‑city projects, while also positioning both firms for export to regions with similar regulatory constraints. Both companies plan to showcase the solution at upcoming industry expos in Seoul and Tokyo.