NERRO (Neuromorphic Edge Real-time Reasoning Operations) is a high-performance AI chip that enables industrial computing.

NERRO is a breakthrough AI hardware technology that enables industrial edge computing in high performance environments. NERRO’s edge computing capabilities deliver unique values for applications in areas of limited or no connectivity, when latency is of high importance, in areas that deal with sensitive information and heightened cyber security risks, or if distributed AI capabilities are critical. The NERRO AI chip is field trainable and can be integrated with existing in sensors and devices that are currently in use.

Oceanit has over 20 years of experience with developing software and hardware systems for AI applications. The NERRO chip is representative of a set of technologies from years of experience in developing sensing technologies for systems with large data rates per sensor (up to many Gbps) with large scale networks to provide “operationally relevant” output in any reasonable amount of time.

Key Features

  • Size less than 4.5mm across
  • Weighs less than 2oz
  • Low power consumption (less than 150mW in active mode)
  • Operates without a network connection (edge computing)
  • Recognizes patterns 50X faster than current industry software

NERRO vs. Cloud Computing


NERRO Chip

Cloud Computing


Field Trainable

Yes

Difficult


Internet Connection

No network required (Edge)

Network connection required


Data Rate

Report on event

Continuous data feed


Latency

Much lower latency due to distributed processing

Latency from communicating to and from cloud servers


Accuracy

More accurate output

Processing delays from ingesting very large data streams and then having to process it all in the cloud


Capital Intensity

No large capital investment for centralized cloud processing

Need large, established infrastructure requiring up-front investment


Security

Lower cyber security risk/easier to maintain control of data

Higher cyber security risk/centralized target


Related products

Privacy Preference Center