The International Business Machines (IBM) is considering using its artificial intelligence chips to lower the costs of operating a cloud computing service, an executive said on Tuesday.
The company is contemplating using a chip called the Artificial Intelligence Unit, as part of its new Watsonx cloud service, said Mukesh Khare, general manager of IBM Semiconductors.
The old version of Watson system faced high costs, which IBM is trying to solve with the improved version, using its chips could lower cloud service costs, as they are very power efficient, Khare added.
The chip is manufactured by Samsung Electronics, which partnered with IBM on semiconductor research, the company has no set date for when the chip could be available for use by cloud customers, Khare stated.
IBM has become one of the big tech companies, along with Google and Amazon, to design its AI chips, but is not trying to design a direct replacement for semiconductors from Nvidia.
IBM’s chips aim to be cost-efficient at inference, which is the process of putting a trained AI system to use in making real-world decisions.
“That’s where the volume is right now. We don’t want to go toward training right now. Training is a different beast in terms of compute. We want to go where we can have the most impact,” said Khare.
IBM has announced the chip in October, but did not disclose the manufacturer or how it would be used.
