Ceva Lands AI Licensing Deal for NeuPro-M NPU with Major U.S. Platform Company
CEVA•Ceva won a licensing agreement for its NeuPro-M neural processing unit with a major U.S. software and AI platform company to underpin a custom AI silicon program targeting next-generation intelligent computing devices. The deal broadens Ceva’s customer base to platform companies and enables OS-to-silicon optimization for power-efficient generative AI inference.
1. Deal Terms and Scope
Ceva has secured a landmark licensing agreement with a leading U.S. software and AI platform company to use its NeuPro-M neural processing unit as the IP foundation for a custom AI silicon program. The collaboration targets next-generation intelligent computing devices by co-designing silicon and software to achieve optimal performance, power and area (PPA) metrics.
2. NeuPro-M Technology Advantages
NeuPro-M provides scalable on-device AI acceleration supporting advanced generative and multimodal inference workloads. Its architecture enables strict power-efficient execution under thermal and battery constraints, while allowing deep OS-to-silicon integration for seamless user experiences.
3. Strategic and Market Impact
This agreement marks one of Ceva’s most strategically significant AI licensing deals, expanding its customer base beyond traditional semiconductor and OEM clients to include major platform companies. It underscores the industry shift toward AI-first computing architectures and strengthens Ceva’s position in the growing custom inference silicon market.
