NTTと日本IBM、オンプレミスAIソリューションを開発

NTT Integration Corporation そして 日本IBM, Ltd. have jointly begun technical verification for the development of a next-generation on-premise AI platform service utilizing “IBM Spyre Accelerator for Power,” the first enterprise-grade AI accelerator in Japan. Through the provision of this service, NTT Integration aims to realize an environment in which companies can utilize AI more safely and efficiently in on-premise environments.

Background: Challenges in the full-scale implementation of AI

In recent years, while the use of generative AI and large-scale AI models has expanded rapidly, new challenges are emerging as we move beyond the PoC and limited application stages and move into the phase of full-scale deployment in business operations.

- 広告

One of these challenges is the increasing power consumption in AI inference infrastructure and the power constraints of data centers. While conventional GPU-based AI infrastructure offers high performance, its high power consumption and installation requirements pose a challenge from the perspective of sustainable AI utilization.

こちらもお読みください： 世界のAIブームから静かに利益を得る日本企業

Furthermore, the introduction and operation of generative AI and RAG (Retrieval Augmented Generation) require advanced expertise, and in many cases, the design and operation of AI infrastructure becomes concentrated in the hands of specific IT personnel. As a result, it has been pointed out that it is difficult for departments to autonomously utilize AI.

Furthermore, for companies handling highly confidential core and business data, there is a strong need to securely utilize AI without moving data to external environments, further increasing the importance of AI utilization in on-premise environments. In particular, in the design departments and production sites of manufacturing companies, highly confidential intellectual property (IP) and confidential information such as design drawings, manufacturing process data, and yield information are strictly managed in on-premise environments, and the desire to utilize this data with AI without taking it outside the company is growing stronger.

ソース PRタイムズ