English 
AI Inside Inc and iNet Corporation have signed a basic agreement to collaborate on the AI integration platform Leapnet and have started discussions toward a proof of concept for a broader initiative called Sovereign Grid. The direction is clear. Move AI inference closer to where data already lives instead of pushing everything to external cloud layers.
Under the agreement, AI Inside will deploy its inference dedicated hardware Cube Atlas 192x along with the Leapnet platform inside iNet’s own data center in Kanagawa. A proof of concept is planned within fiscal year 2026 to test how this setup performs in a real operating environment.
The larger idea behind Sovereign Grid is not a single centralized AI buildout. It is a distributed model where domestic data center operators participate in building an AI inference network across Japan. This reflects a shift in how data centers are being positioned. Not just storage facilities anymore, but AI factories that convert electricity into usable intelligence.
こちらもお読みください： テンチジン、ALOS-4データマーケットプレイスを立ち上げました
iNet, which operates its Dream Cloud service, is positioning itself for this shift by strengthening its AI business and moving toward a next generation AI data center model. With this collaboration, it aims to offer services like generative AI infrastructure, AI OCR, and internal document search directly from its own facilities, keeping data within Japan for security sensitive workloads.
について AI Inside, the partnership is a starting point. The plan is to expand similar deployments across more domestic operators and gradually build a national level inference network powered by Leapnet and Sovereign Grid.
日本語