Lattice Semiconductor has released an updated version of its sensAI software stack, seeking to strengthen its position in low-power Edge artificial intelligence as competition intensifies among chipmakers targeting industrial and automotive systems.
The US-based group said the latest iteration of sensAI expands support for artificial intelligence models, improves performance per watt, and simplifies deployment for developers building applications at the Edge of networks, where power efficiency and responsiveness are critical.
Edge AI, which processes data locally rather than in centralised data centres, has become a focus for manufacturers as factories, vehicles, and consumer devices adopt more autonomous and real-time capabilities. Lattice, which specialises in low-power programmable chips, is positioning sensAI as a toolset that allows engineers to integrate AI functions without the energy and cost penalties associated with more power-hungry processors.
Raemin Wang, Vice President, Segment Marketing at Lattice Semiconductor, said demand was rising for systems that combine low power consumption with higher levels of intelligence. He said the updated sensAI stack was designed to help customers deploy AI more quickly across applications ranging from industrial automation to vehicle infotainment and security systems.
Version 8.0 of sensAI adds new purpose-built models for areas such as human-machine interfaces, multi-object detection, and defect detection, alongside upgraded compiler tools and broader support for different AI network topologies. Lattice said these changes would allow its field-programmable gate arrays to deliver higher performance while maintaining industry-leading power efficiency.
The company has also emphasised ease of use, introducing a more complete toolchain with Python application programming interface integration, as well as greater flexibility in deployment. Features include a simplified RISC-V codebase and YAML-based automation intended to speed up prototyping and customisation.
Lattice is working with industrial partners to apply the technology in factory environments. Naoki Nakamura, General Manager of the Numerical Control System Department at Mitsubishi Electric, said collaboration with Lattice would combine FPGA-based AI acceleration with industrial expertise to develop scalable and secure edge AI solutions for next-generation automation equipment.
The move comes as semiconductor groups increasingly target Edge computing as a growth area, betting that demand for localised AI processing will expand as industries seek to reduce latency, bandwidth usage, and energy consumption.
