Nvidia upgrades flagship chip to handle bigger AI systems

FILE PHOTO: NVIDIA HGX AI Supercomputer on display during the annual Foxconn Tech Day in Taipei

By Stephen Nellis

(Reuters) - Nvidia on Monday added new features to its top-of-the-line chip for artificial intelligence, saying the new offering will start to roll out next year with Amazon.com, Alphabet's Google and Oracle.

The H200, as the chip is called, will overtake Nvidia's current top H100 chip. The primary upgrade is more high-bandwidth memory, one of the costliest parts of the chip that defines how much data it can process quickly.

Nvidia dominates the market for AI chips and powers OpenAI's ChatGPT service and many similar generative AI services that respond to queries with human-like writing. The addition of more high-bandwidth memory and a faster connection to the chip's processing elements means that such services will be able to spit out an answer more quickly.

The H200 has 141-gigabytes of high-bandwidth memory, up from 80 gigabyte in its previous H100. Nvidia did not disclose its suppliers for the memory on the new chip, but Micron Technology said in September that it was working to become an Nvidia supplier.

Nvidia also buys memory from Korea's SK Hynix, which said last month that AI chips are helping to revive sales.

Nvidia on Wednesday said that Amazon Web Services, Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure will be among the first cloud service providers to offer access to H200 chips, in addition to specialty AI cloud providers CoreWeave, Lambda and Vultr.

(Reporting by Stephen Nellis in San Francisco. Editing by Sam Holmes.)