HPE is rolling out a cloud computing service designed to power artificial intelligence systems similar to ChatGPT.
HPE’s service is being used by a few customers now, with more availability in North America by year’s end and in Europe next year.
The move puts HPE into direct competition with cloud computing providers such as AWS, Microsoft and Google, all of whom are racing to adapt their massive data centres for an era in which chatbots, image generators and other AI-backed services are drawing hundreds of millions of users.
That shift toward AI is shaking up the cloud computing market because data centres must be built very differently to handle such work.
In a typical cloud computing data center, software is used to chop up a single physical server into many smaller virtual machines that can then be rented out to customers.
But data centres for artificial intelligence take an opposite approach.
Those systems aim to link together hundreds or even thousands of computers and make them look like one giant computer.
HPE has been developing that kind technology to weave together computers for years for systems like the Frontier supercomputer, which HPE developed with Oak Ridge National Laboratory in the United States and which is currently the fastest computer in the world.
Justin Hotard, executive vice president and general manager of HPE’s high-performance computing and artificial intelligence unit, said the company will use its experience in supercomputers to offer a service specifically for what are called large language models, the technology behind services like ChatGPT.
“We see it as complementary and very different than what our fellow cloud partners provide. It’s not trivial or freely accessible,” Hotard said in a interview.
About Author
