IBM considers using its own AI chips in new cloud services to reduce costs

CTOnews.com IBM is considering using artificial intelligence chips designed in-house to reduce the cost of cloud computing services.

The company is considering a chip called the AI Unit as part of its newly launched Watson X cloud service, general manager Kumar Carey told Reuters in an interview at the Semiconductor Conference in San Francisco.

IBM wants to make it easier for users to use generative AI technology that can learn human text and language better than Watson, its first major AI system launched more than a decade ago.

It is said that one of the main obstacles currently facing the old Watson system is high cost, and IBM hopes to solve this problem with its own AI chip this time. Carey said using his own chips can reduce the cost of cloud computing services because they are very energy efficient.

IBM introduced its AI chips in October last year, but did not disclose details such as the manufacturer or how they are used. The chip is manufactured by Samsung Electronics, which has a partnership with IBM in semiconductor research, and IBM is considering it for Watson X, Carey said.

IBM has yet to determine when its chips will be available for cloud customers, but Carey said the company already has thousands of prototype chips in operation.

CTOnews.com previously reported that other tech giants, such as Alphabet's Google and Amazon, have begun designing their own AI chips.

But Carey said IBM is not trying to design a direct replacement for Nvidia chips, which are leading the way in training large AI systems with large amounts of data. Instead, IBM's chips are designed to be cost-effective in what industry insiders call inference-the process of using well-trained AI systems for actual decision-making.