EdgeCortix SAKURA-I AI Accelerator Demonstrates Robust Radiation Resilience, Suitable for Many Orbital and Lunar Expeditions.

TOKYO--( BUSINESS WIRE )-- EdgeCortix® Inc. , a leading fabless semiconductor company specializing in energy-efficient Artificial Intelligence (AI) processing at the edge, today announced that its SAKURA-I AI Accelerator is suitable for space missions including in Earth orbit and on the lunar surface, demonstrating high levels of radiation resiliency. NASA Electronic Parts and Packaging Program (NEPP) executed two phases of testing on EdgeCortix’s AI accelerator, subjecting it to both proton and heavy ion radiations to evaluate its ability to withstand radiation impact in space-like environments.

The NEPP testing initiative was commissioned to advance the goal of achieving full autonomy in space. While recent advancements in machine learning and computer vision have shown significant promise in enhancing sensor processing capabilities, the computational demands of these algorithms often exceed the capabilities of many embedded central processing units (CPUs). Additionally, graphics processing units (GPUs), though powerful, typically consume over 40 watts—far exceeding the acceptable power thresholds for most space missions. To address these critical challenges, NEPP evaluated several commercially off-the-shelf (COTS) low-power AI accelerators, including EdgeCortix’s SAKURA-I, as a potential solution.

Proton and heavy ion testing, conducted at Massachusetts General Hospital and Lawrence Berkeley National Labs, confirmed that EdgeCortix’s accelerator demonstrates radiation tolerance suitable for environments such as low Earth orbit, geosynchronous orbit, and the lunar surface. The tests revealed that SAKURA-I experienced no destructive events, while transitory radiation effects occurred at a significantly lower rate compared to other similar products tested. These results underscore the potential to deploy COTS edge AI accelerators in space, paving the way for more complex, efficient, and cost-effective operations in near-term missions.

“The radiation resilience demonstrated by SAKURA-I marks a significant milestone in the advancement of AI-driven space exploration,” said Sakyasingha Dasgupta, CEO and Founder of EdgeCortix. “Processing billions of data points swiftly and accurately at the ultimate edge—space—without relying on Earth-based systems for analysis and decision-making was once considered out of reach. NASA’s testing confirms that high-performance, autonomous data processing during spaceflight is no longer just a possibility but a reality. EdgeCortix stands ready to deliver cutting-edge AI solutions that empower autonomous space applications today.”

Details on the proton and neutron test results can be found in the published papers below:

EdgeCortix SAKURA-I Machine-Learning, PCIe Accelerator SEE Heavy Ion Test Report
https://ntrs.nasa.gov/citations/20240015800

EdgeCortix SAKURA-I Machine-Learning, PCIe Accelerator SEE Proton Test:
https://ntrs.nasa.gov/citations/20240006221

About EdgeCortix

Pioneering the future of the connected intelligent edge, EdgeCortix is a fabless semiconductor company focused on energy-efficient AI processing of Generative-AI workloads at the edge. Founded in 2019 with R&D headquarters in Tokyo, Japan, EdgeCortix delivers a software-first approach with its patented “hardware and software co-exploration” system to design an AI specific runtime reconfigurable accelerator from the ground up. EdgeCortix’s products disrupt the rapidly growing edge AI hardware markets including defense, aerospace, smart cities, industry 4.0, autonomous vehicles and robotics.

For more information about EdgeCortix, visit https://www.edgecortix.com/en/ .

About NASA Electronic Parts and Packaging Program (NEPP)

The NEPP Program generates technical knowledge and recommendations about electrical, electronic, electromechanical (EEE) part performance, application, failure modes, test methods, reliability and supply chain quality within the context of NASA space flight missions and hardware. This information is made available to the NASA and high-reliability aerospace community through publications, web pages and links published on https://nepp.nasa.gov/ .

View source version on businesswire.com: https://www.businesswire.com/news/home/20250127424314/en/