Ubitus has announced its official selection for Japan’s Ministry of Economy, Trade and Industry (METI) Large-Scale Growth Investment Grant Program for Medium-Sized and Small Enterprises. The company reports it will invest JPY 17 billion in expanding regional distributed GPU infrastructure, leveraging NVIDIA’s Blackwell GPU architecture to enhance its NeoCloud platform. This effort aims to advance Japan’s high-performance computing and support the nation’s growing artificial intelligence (AI) requirements.
The approved project, “Regional Distributed GPU Infrastructure Development for the Generative AI Era,” specifically targets domestic AI infrastructure expansion. Ubitus says its NeoCloud platform will implement layered scheduling, distributed computing, and elastic resource management. These capabilities are designed to enhance computational efficiency, minimize latency, and dynamically scale compute resources to address regional demand.
Ubitus claims the next-generation NeoCloud architecture will enhance Japan’s data sovereignty, promote local deployment of AI workloads, and reduce energy and bandwidth bottlenecks that are often found in centralized facilities. The architecture is intended to build resilience and enable a flexible, scalable nationwide GPU foundation.
The company notes prior collaboration through the “GENIAC” Program, which included development of a 405-billion-parameter East Asia large language model supporting Japanese, Traditional Chinese, and English. Benchmarking reportedly showed this model surpassed GPT-4 in selected knowledge-based tests. Ubitus has also deployed advanced AI language models for applications such as tourism, with “Chokimaru,” an AI guide in Maizuru City, Kyoto, enabling real-time multilingual Q&A and translation via browser and mobile devices.
Ubitus states that it will continue to strengthen technical and industrial partnerships within Japan, focusing on distributed AI infrastructure for data center operators, governments, and enterprises. Key operational targets include reducing latency and energy bottlenecks, improving localized compute, and supporting applications in industries ranging from culture and tourism to healthcare and enterprise AI deployments.
Source: Ubitus
