Neuromorphic Chip Maker Taps Synopsys for Scale
Innatera, a company developing brain-inspired neuromorphic chips for edge AI, has selected Synopsys' simulation technology to help design its processors. The move aims to accelerate the development of low-power AI chips for wearables, smart home devices, and industrial digital twin sensors.
Designing novel processors like neuromorphic chips presents immense complexity. Companies use Electronic Design Automation (EDA) software from leaders like Synopsys, which increasingly integrates AI to automate and optimize chip design, aiming to reduce iterations, improve power-performance-area (PPA) objectives, and accelerate time-to-market. Netherlands-based Innatera, a 2018 spin-off from Delft University of Technology, has developed its processors based on Spiking Neural Networks (SNNs). This brain-inspired, event-based architecture allows for extreme efficiency, with the company claiming its chips can be 100 times faster and use 500 times less energy than conventional processors for pattern recognition tasks. Innatera's "Pulsar" chip features a hybrid architecture, combining the primary SNN engine with a traditional RISC-V microcontroller and a Convolutional Neural Network (CNN) accelerator. This design allows it to function as a sole, all-in-one processor for a sensor, handling everything from data conditioning to complex AI inference directly on the device. For robotics and industrial automation, this technology enables "always-on" sensing and real-time decision-making with sub-milliwatt power consumption. This is critical for battery-powered systems like autonomous drones or mobile industrial robots, where low latency and energy efficiency are paramount for navigation and object recognition without cloud reliance. The neuromorphic computing market is projected to grow from tens of millions in 2024 to over $1.2 billion by 2030. Innatera competes with established players and startups, including Intel's Loihi 2 research chip, IBM's NorthPole, and
