Air Force Validates SAKURA-II Edge AI Chip for Aerospace and Defense Applications

The U.S. Air Force has validated the performance of an edge AI computing chip developed by a Japanese company. (Image: EdgeCortix)

July 7, 2026

The U.S. Air Force has completed a breakthrough demonstration of an edge artificial intelligence (AI) processing chip in an exercise supporting advanced multi-domain battlefield connectivity and AI-enabled mission systems.

The demonstration validated the ability of the EdgeCortix SAKURA-II platform to deliver high-performance, low-power AI inference capabilities in operationally relevant domains including aerial operations while supporting both commercial and custom defense AI applications. EdgeCortix is a Japanese fabless semiconductor company that specializes in energy-efficient AI processing for edge applications.

“The U.S. Air Force and EdgeCortix worked together to integrate SAKURA-II into a relevant mission system and fly it in a large force exercise, validating AI inference in flight with a tactically relevant application in operationally relevant scenarios,” said Lt. Col Spencer Liedl, KC-135 Operational Test Director, Roland R. Wright ANGB, UT, Air National Guard Air Force Reserve Command Test Center (AATC).

As part of the DIU project, EdgeCortix successfully completed:

AI Benchmarking and Performance Validation: Activities included independent validation support from the Carnegie Mellon University Software Engineering Institute (CMU SEI).
Radiation Resilience Testing of the SAKURA-II AI Accelerator Platform for Orbital and Lunar Missions: NASA heavy ion testing validated high radiation resiliency with no destructive events and minimal transient effects, supporting deployment in space and defense AI applications from LEO through cislunar regimes.
Integration and Flight Testing: Included a prototype Advanced Intelligent Gateway System designed to provide resilient multi-domain battlefield connectivity across joint and coalition forces.

EdgeCortix’s SAKURA-II edge computing chip is a dual64-bit LPDDR4x AI accelerator with a performance of 60 TOPS (INT8) and 30 TFLOPS (BF16), 68 GB/sec of DRAM bandwidth and 20 MB of on-chip SRAM. (Image: EdgeCortix)

The prototype system demonstrated the ability to execute advanced AI workloads efficiently at the tactical edge under demanding environmental and operational conditions. The radiation validation activities also built upon prior EdgeCortix work with NASA validating the resilience of the SAKURA-II AI accelerator platform for orbital and lunar mission environments .

“This milestone validates the readiness of the EdgeCortix SAKURA-II platform for demanding aerospace and defense environments,” said Dr. Sakyasingha Dasgupta, Founder and CEO of EdgeCortix. “The successful benchmarking, radiation testing, airborne integration, and operational flight demonstration of SAKURA-II highlight the role of energy-efficient edge AI platforms in enabling trusted autonomy and resilient AI inference across next-generation defense and space systems.”

EdgeCortix developed the SAKURA-II based on its "Dynamic Neural Accelerator" (DNA) technology , which the company describes as a "flexible, run-time reconfigurable, highly-parallelized and efficient architecture." DNA is a "highly-efficient and powerful neural network IP core that can be paired with any host processor. Achieving exceptional parallelism and efficiency through run-time reconfigurable interconnects between compute elements," according to the EdgeCortix website.

The DIU Success Memorandum recognizes the validation of commercially developed edge AI technology in operationally relevant defense environments and establishes a foundation for potential future transition, deployment, and follow-on production opportunities supporting next-generation aerospace and defense missions. The successful flight testing activities further increased the Technology Readiness Level (TRL) of the EdgeCortix platform for airborne deployments.

Source