Ceva’s Strategic Play in the AI Surge: From Smart Edge to Scalable AI Platforms

Ceva, Inc. is known as a leading licensor of silicon and software IP for connected devices and is actively transforming its business to capture growth from the global AI boom. Rather than relying solely on traditional wireless connectivity IP, the company has strategically expanded its portfolio, partnerships, and ecosystem around embedded AI processors and related software tools, positioning itself at the intersection of edge computing, AI inference, and smart device intelligence.

AI Drives New Licensing Momentum

One of the clearest indicators of Ceva’s AI advancement is its growing contribution of AI-related revenue. In 2025, Ceva reported that about one-third of its licensing revenue came from AI processor deals, marking the first time AI has taken such a meaningful share of its licensing mix. These AI agreements often come with higher royalty potential than traditional IP agreements, which enhances the company’s long-term revenue prospects as AI adoption expands across devices.

This shift reflects a broader strategic realignment: Ceva is no longer simply a provider of connectivity and sensor IP, but a supplier of AI inference engines and DSP accelerators that enable smart features, from vision and speech to contextual awareness.

NeuPro NPU Family Leads the AI Charge

At the core of Ceva’s AI strategy is its NeuPro™ family of Neural Processing Units (NPUs) scalable AI processors designed for efficient inference on edge devices such as microcontrollers, IoT endpoints, wearables, and sensor fusion systems. These NPUs are optimized for low-power operation and high performance per watt, which is critical for smart edge applications where cloud connectivity, latency, and energy usage are major constraints.

The NeuPro-Nano variant has received industry recognition for excellence in edge computing, winning a 2025 IoT Edge Computing Excellence Award from IoT Evolution World. This award underscores the role of NeuPro-Nano as a cornerstone of physical AI, embedding real-time intelligence within devices where latency, privacy, and reliability matter most.

Ecosystem Partnerships Extend Market Reach

Ceva understands that IP alone does not create market pull; developers need tools, models, and integrations that accelerate time to market. To that end, the company has built a robust embedded AI ecosystem through strategic partnerships:

Cyberon and AIZIP provide pre-optimized neural network models (e.g., keyword spotting, face recognition), which simplify deployment on NeuPro platforms.

Edge Impulse integration with NVIDIA’s TAO Toolkit helps developers train, retrain, and deploy AI models with minimal coding, accelerating development cycles and broadening use cases.

These collaborations reduce barriers to entry for customers and broaden Ceva’s addressable market by enabling richer AI functionality in smart watches, IoT endpoints, and other embedded systems.

Strategic OEM and IP Licensing Wins

Ceva’s AI strategy is also reflected in key industry partnerships. A major engagement with Microchip Technology will see Ceva’s NeuPro NPUs integrated across Microchip’s product lines — from compute and communication to security and infrastructure products. This reinforces Ceva’s reach across industrial, automotive, and data center segments, beyond purely consumer electronics.

In addition, Ceva’s financial reporting highlights multiple AI DSP agreements with OEMs in home appliances and automotive spaces, including advanced driver assistance systems (ADAS). The company even noted engagements within automotive chiplet architectures, strengthening its foothold in safety-critical and performance-demanding environments.

Broadening Relevance in the Smart Edge Era

Ceva’s evolution reflects broader industry trends: AI is shifting from cloud-centric compute to distributed edge inference, where on-device intelligence delivers lower latency, greater privacy, and improved power efficiency. By pairing scalable NPUs with a comprehensive SDK “NeuPro Studio” Ceva is addressing both hardware and software needs of AI developers, a combination that’s crucial for adoption.

This is evident in the company’s device footprint: Ceva-powered silicon has surpassed 20 billion shipped devices, a milestone that highlights its deep legacy in IoT and its emerging role in the next generation of AI-enabled products.

Bottom line: Ceva’s strategy to capitalize on the AI surge is multifaceted: it involves expanding its core AI processor IP, fostering a developer-friendly ecosystem with partners, and securing licensing wins that embed its technology across a wide range of smart devices. By aligning its IP portfolio with the growing demand for **edge AI inference — and by providing the tools and partnerships that make AI deployment practical — Ceva is positioning itself not just as an IP vendor but as a key enabler of physical AI in the next wave of connected intelligence.

CONTACT CEVA

Also Read:

Ceva-XC21 Crowned “Best IP/Processor of the Year”

United Micro Technology and Ceva Collaborate for 5G RedCap SoC and Why it Matters

Ceva Unleashes Wi-Fi 7 Pulse: Awakening Instant AI Brains in IoT and Physical Robots