Fujitsu has announced the Fujitsu Cloud Service Generative AI Platform, a secure, cloud-based solution designed to support enterprises handling sensitive data. The platform, set for release in Japan during fiscal year 2025, will provide businesses with private and compliant AI-driven workflows while preventing data leakage and unintended AI learning. A trial program for the service will begin in Japan on February 13, 2025, with a global rollout planned for the future.
The platform integrates Fsas Technologies Inc.’s Private AI Platform on PRIMERGY, Supermicro’s GPU servers, and Fujitsu’s Takane large language model (LLM)—a key technology from the Fujitsu Kozuchi AI service, known for its Japanese language proficiency. By hosting AI processing in Fujitsu’s data centers, the service ensures confidential data remains within private cloud areas, addressing compliance and security concerns. Fujitsu also plans to expand support for additional third-party LLMs to meet diverse enterprise needs. The new offering will be added to Fujitsu’s Computing as a Service (CaaS) portfolio, with end-to-end support for businesses adopting generative AI.
Key Features of Fujitsu’s Generative AI Platform:
• Private data management – Fine-tuned models and retrieval-augmented generation (RAG) data are stored in dedicated private cloud areas to ensure security and compliance.
• Cost-effective shared GPU resources – Shared GPU infrastructure lowers per-hour costs compared to on-premise AI environments.
• Secure AI access – Fujitsu’s AI security enhancement technology prevents unauthorized data access and includes vulnerability assessments and third-party audits.
• Enterprise-grade cloud security – Hosted in Fujitsu’s JDCC Tier 4 equivalent data center, offering high-security cloud solutions for AI workloads.
“With the growing demand for secure and efficient AI solutions, Fujitsu’s new generative AI platform ensures enterprises can leverage AI while safeguarding their sensitive data,” the company stated.
