NXP to buy US edge AI chip startup Kinara for $307m

NXP Semiconductors is to buy US edge AI chip startup Kinara for $307m in cash.

Kinara was a spinout of Stanford University in 2014 called Deep Vision and developed low power neural network processors that can run many types of AI model at the end of the network, including multi-modal generative AI models. The deal is is expected to close in the first half of 2025, subject to regulatory clearances.

NXP is an existing partner with Kinara to link its microcontrollers and connectivity chips to the Ara-1 and Ara-2 discrete NPUs for applications in vision, voice and gesture recognition, starting with smart cameras. Both devices are programmable, mapping of the inference graphs onto proprietary neural processing units for maximizing edge AI performance. This allows the chips to run a range of AI models, from convolutional neural networks to transformer-based generative AI models as well as  agentic AI in the future.

• Dennis Segers joins Kinara board

Kinara is working with Mirasys in India on vision applications such as smart cameras using the first generation Ara-1, while the Ara-2 provided more performance of 40 TOPS for generative AI models with up to 7bn parameters in a power envelope of 6W.

Kinara also provides a software development kit for designers to optimize AI model performance and streamline the deployment along with the model libraries and model optimization tools. These will all be integrated into NXP’s eIQ AI/ML software development environment.

“The industrial market is going through a transformation, with generative AI helping to deliver major improvements in efficiency, sustainability, safety and predictability, and in many instances, unlock new use cases and functionality,” said Rafael Sotomayor, executive vice president and general manager, Secure Connected Edge at NXP. “Adding Kinara’s AI capabilities to our broad intelligent edge portfolio creates a scalable platform for new classes of AI-powered systems. Together, we can help our customers simplify complexity and accelerate time to market as they create transformative AI systems.”