Soldiers could soon operate a rugged AI system at the frontline, enabling real-time analysis and decision-making without relying on cloud connectivity.
Under a roughly $2-million Phase II award from the US Army’s xTech Small Business Innovation Research program, GSI Technology will develop a field-ready edge AI platform, building on earlier phase work that validated its core approach.
The system will be built around GSI’s Gemini-II processor, which uses a compute-in-memory architecture to process data directly where it is stored, reducing latency and power consumption.
Designed for size, weight, and power constraints, the platform is intended to enable real-time processing of battlefield data, including sensor inputs and targeting information, at the point of need.
“This award represents an important step toward field deployment in defense applications,” said Lee-Lean Shu, chairman and CEO of GSI Technology.
“Gemini-II is purpose-built to deliver high-performance AI within tight power and space constraints, enabling autonomous decision-making at the point of need without reliance on cloud or reachback infrastructure.”
The Phase II award builds on an earlier contract of up to $250,000 that validated Gemini-II integration with army AI models and identified suitable edge algorithms, as the Pentagon pushes ahead with an “AI-first” approach to deployment.
