Delft University of Technology spin-off's Spiking Neural Processor utilizes a proprietary analog-mixed signal computing architecture that harnesses spiking neural networks to deliver unprecedented, brain-inspired cognition performance within an ultra-low power envelope.

LAS VEGAS, Jan. 7, 2025 -- Innatera, the pioneer of groundbreaking ultra-low power neuromorphic processors, is showcasing its Spiking Neural Processor (SNP) that transforms the way battery-powered devices make sense of the physical world at CES 2025. Demonstrations will take place January 7-10 at the Venetian North Tower Suite #26-136.

"At this pivotal moment in computing, Innatera's breakthrough Spiking Neural Processor delivers unmatched energy-efficient, brain-inspired cognition for sensors, unlocking the promise of ambient intelligence," said Sumeet Kumar, CEO of Innatera. "This revolutionary processor provides an all-in-one solution that simplifies and optimizes sensor data processing at the edge."

Innatera's SNP combines a Spiking Neural Network (SNN) engine with a RISC-V processor core and other accelerators to deliver a complete solution with exceptional performance in energy-constrained environments. The single-chip solution brings intelligence closer to sensors, enabling next-generation AI and signal processing for applications in consumer electronics, smart homes, and industrial IoT, such as audio interfaces, touch-free interfaces, presence detection, activity recognition, and ECG recognition.

The SNP achieves high-performance pattern recognition at the sensor edge and enables real-time analysis of sensor data to detect and identify embedded patterns, with sub-milliwatt power dissipation and sub-millisecond latency.

Ambient Intelligence marks a major departure from computing technology as we know it, paving the way for a future where digital interactions are as natural as breathing.

At CES 2025, Innatera is demonstrating how the SNP can transform computing in several real-world applications:

• Audio Scene Classification: Audio scene classification allows devices to be aware of the environment they operate in and use this information to adapt their operation. For example, in noise canceling headphones, this might mean adapting to the ambient noise profile of an airplane or a city bus.
• Robust Human Presence Sensing: The detection of human presence is important in a wide range of indoor and outdoor applications, such as security cameras, smart lighting, video doorbells and smart TVs. Using a radar sensor, this demo showcases always-on, privacy-preserving human presence detection with exceptional accuracy and power efficiency.
• Robust People Counting Using Far Infrared Sensors: Innatera will showcase how its SNP enables advanced people counting and human presence detection with passive infrared sensors. Infrared technology is a non-intrusive, low-light, and privacy-preserving method for people counting and human presence detection.
  
  

The company's presence at CES 2025 follows a remarkable year of growth and development for the innovation-driven Delft University of Technology spin-off. Earlier this year, the company announced the oversubscription of a Series A $21-million funding round that is now accelerating the company's development of neuromorphic processors.