Oracle readies its databases for the AI wave

AI

Oracle Corp. is declaring itself a competitor alongside major hyperscalers as companies look for large-scale databases that can meet the demands of artificial intelligence.

“The business has progressed from those mission critical types of workloads, whether it’s financials or supply chain or transactions, into the AI space,” said Leo Leung (pictured), vice president of OCI and technology at Oracle. “We’ve had a great few years running these big AI workloads for companies that are training these new models.”

Leung spoke with theCUBE’s chief analyst, Dave Vellante, at the Nvidia GTC event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how Oracle is making itself essential to the AI space and the evolution of data infrastructure.

Structuring databases for AI workloads

Oracle focuses on supporting and simplifying database infrastructures for companies, which can allow them to better implement AI. The tech giant’s enterprise stack has AI services embedded within its applications, which gives its customers more flexibility, according to Leung.

“A lot of the world’s companies store their … critical data inside of Oracle Database, and now they want to be able to do AI on that data,” said Leung, who added that many companies are interested in providing internal proprietary data to its models. “We see that as the next trend, and … our position as a database company, as a cloud company for critical workloads puts us, I think, in a really good position to go address that.”

Oracle recently announced an expansion of its partnership with Nvidia Corp., which has been dominating the market for graphic processing units that can support AI models. However, not all customers will require the same level of computing power, believes Leung.

“The very, very large scale consumer-facing types of inferencing products … There’s going to be a different scale of that versus an enterprise, which is going to have a few thousand end customers, 10,000 end customers,” he said.

Although Oracle remains what Vellante calls “the database king,” he questioned if it was a hyperscaler like Amazon Web Services or Google. Leung had a response.

“We actually have more regions and more presence in more places than anybody else,” he said. “We’ve got customers like Uber that are running on us … You can’t serve a customer like that unless you’re a hyperscaler.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of the Nvidia GTC event: 

Photo: SiliconANGLE