NeuPro-M selected as NPU IP foundation for custom AI silicon program, enabling OS-to-silicon optimization for next-generation intelligent computing devices
Ceva, Inc., a leading licensor of smart edge silicon and software IP, announced that it has entered into a landmark AI licensing agreement with a leading U.S. software and AI platform company for a custom AI silicon program for next-generation intelligent computing devices. This agreement expands Ceva’s customer base beyond traditional semiconductor companies and device OEMs to software platform companies that are increasingly designing custom silicon to optimize performance, power, area (PPA) and overall user experience.
“The decision by one of the industry’s leading software and AI platform companies to build custom AI silicon on NeuPro-M reflects a broader shift to AI-first computing architectures,” said Amir Panush, CEO of Ceva. “Intelligent devices are increasingly expected to sense, reason, and act locally, driving demand for AI acceleration that delivers high performance within tight power and thermal constraints. As AI workloads are increasingly distributed across clouds and edge devices, platform companies are optimizing the entire stack, from silicon and software frameworks to operating system integration and user experience. We mark this as one of the most strategically important AI licensing agreements in Ceva’s history. This reflects the growing role of AI acceleration in shaping the future of computing.”
Also read: AiThority interview with Matej Bukovinski, Chief Technology Officer at Nutrient
Leading technology platform companies are increasingly recognizing that custom AI silicon is essential to optimize performance, power efficiency, and full-stack control at scale. For companies that own both the operating system and the hardware platform, co-designing silicon and software creates a critical advantage. Particularly for portable edge computing devices with severe thermal and battery constraints, enhanced OS-to-silicon optimizations enable performance and power efficiency improvements not possible with off-the-shelf processors. Just as the CPU defined general-purpose computing and the GPU accelerated graphics and parallel workloads, AI acceleration is emerging as the third foundational layer of the computing stack, driving a new generation of custom inference silicon and positioning the NPU as a core architectural element of future intelligent computing platforms.
Customers selected NeuPro-M to provide scalable, power-efficient AI acceleration for advanced on-device inference workloads. This architecture enables efficient execution of generative AI, multimodal AI, emerging agent AI workloads, and other machine learning applications while operating within the power, area, and thermal constraints of intelligent edge computing devices. NeuPro-M allows customers to integrate advanced AI capabilities directly into custom silicon architectures, providing the flexibility to co-optimize performance, power efficiency, and user experience across the entire hardware and software stack. As part of the program, Ceva worked closely with customers to implement advanced neural network optimizations tailored to their target AI workloads to further improve inference efficiency and performance.
