Introducing RAIDEN: The Physical AI compute platform that adapts as models change and scales as systems grow
RAIDEN™ unleashed our next-generation, energy-efficient AI chiplet platform for systems that sense, reason, act and scale in the real world. Scaling from a single die to a four-die configuration, RAIDEN unifies compute, memory, and connectivity under one architecture and software stack. RAIDEN-X4: up to 3.36 PFLOPS FP4 and 256 GB of memory, with customer design wins secured.
Why does Physical AI need to scale out? EdgeCortix Founder and CEO Dr. Sakya Dasgupta shares the thinking behind RAIDEN and explains how chiplets can help distribute intelligence across machines, factories, networks, and autonomous systems.
Nikkei Asia: Japan's EdgeCortix looks to work with SpaceX on space data center chips, Edge AI startup unveils new product, signs multimillion-dollar deal with Kawasaki
Nikkei Asia recently sat down with our Founder and CEO, Dr. Sakya Dasgupta, to discuss RAIDEN and what’s next for EdgeCortix. The article highlights RAIDEN’s scalable, energy-efficient architecture for Physical AI applications across robotics, manufacturing, aerospace, defense, telecom and edge infrastructure, including our multimillion-dollar agreement with Kawasaki Heavy Industries, and why it could ultimately take Physical AI compute from the factory floor to orbit.
Bringing Data Center AI Performance to the Edge at "Innovate with TSMC!" 2026
How can advanced AI run closer to where data is created? At the TSMC 2026 North America Technology Symposium, Founder and CEO Dr. Sakya Dasgupta shares EdgeCortix’s vision for energy-efficient edge AI. Discover how MERA™, SAKURA-II, and our next-generation chiplet architecture support the move of generative AI, large language models, and Physical AI into edge applications.
EdgeCortix Signs Multi-Year, Multimillion-Dollar Agreement with Kawasaki Heavy Industries to Advance Next-Generation AI-Enabled Aerial Defense Systems
EdgeCortix and Kawasaki Heavy Industries have entered into a multi-year Agreement to advance AI-enabled aerospace and defense systems. The collaboration pairs Kawasaki's aerospace and systems expertise with EdgeCortix's next-generation chiplet-based AI computing platform and MERA software, with an initial program value of several million U.S. dollars through 2028. The companies will develop and integrate technology bringing greater intelligence, adaptability and energy efficiency to future mission-critical aerial defense systems.
In this episode of AI with Sally, EdgeCortix Founder and CEO Sakya Dasgupta joins Sally Ward-Foxton to discuss the company’s software-first approach to AI inference, the architecture behind SAKURA-II, and the challenge of delivering high-performance, energy-efficient AI at the edge. The conversation also explores EdgeCortix’s work in space and aerospace, including NASA-related applications, Japan’s semiconductor resurgence, and what’s next for the company’s next-generation chiplet platform.
Forbes: “From Cloud AI To Real-World Agents: The Shift To Intelligence That Acts”
As AI agents move into robots, factories, drones, and satellites, energy efficiency and reliable local processing become essential. In his Forbes Technology Council article, Dr. Sakya Dasgupta explores the shift toward intelligence that acts and why useful decisions per joule, real-world latency, and system reliability are increasingly important measures of AI performance.
