AI optimization software company Nota and edge AI semiconductor company Syma AI have partnered to target the physical AI market. In response to the growing demand for on-device AI that runs in real-time at industrial sites, the two companies plan to jointly develop and commercialize a solution that combines lightweight AI technology with low-power semiconductors.

The two companies announced that they recently signed a strategic partnership at the Cima AI headquarters in San Jose, California. The agreement includes the joint development of on-device AI solutions, expanded technological cooperation, identification of clients, and joint execution of pilot projects.

This collaboration aligns with the technical requirements arising from the proliferation of physical AI. AI operating in real physical environments, such as robots, mobility systems, and industrial automation equipment, must simultaneously ensure fast inference performance and power efficiency. As reliance on the cloud decreases and immediate decision-making and operation are required on-site, the simultaneous optimization of software and hardware is becoming increasingly important.

Nota plans to support the efficient operation of models on Syma AI's NPU-based MLSoCs through Netspresso, an AI model lightweighting and optimization platform. According to the company, Netspresso is a platform capable of compressing, optimizing, and deploying models to suit various hardware environments. It focuses on securing AI computational performance even in power-constrained environments by significantly reducing model size while minimizing accuracy degradation.

Syma AI provides MLSoCs and software development environments for edge environments. The company's Modalix MLSoC is a system-on-chip designed for both multimodal inference and energy efficiency, while the Palette SDK is introduced as a tool to simplify the deployment process of complex edge AI applications. Both companies plan to enhance technical compatibility from the development stage by linking the Netspresso SDK with the Syma AI development environment.

Applications are expected to materialize first in the fields of intelligent transportation systems, safety, and security. Nota plans to optimize NVA, a generative AI-based video surveillance solution, for Syma AI hardware and apply it to industrial sites; based on this, the company intends to expand the scope of cooperation across the entire spectrum of physical AI, including robotics and mobility. This collaboration can be viewed as an attempt to coordinate software and semiconductor technologies to realize AI systems that are practically usable in edge environments.