FILE PHOTO: A smartphone with a displayed NVIDIA logo is placed on a computer motherboard in this illustration taken March 6, 2023. REUTERS/Dado Ruvic/Illustration
OAKLAND, California (Reuters) - Snowflake a cloud data analytics company, is partnering with computing company Nvidia to allow customers ranging from financial institutions to healthcare and retail to build AI models using their own data.
The two companies announced the partnership at Snowflake Summit 2023 on Monday.
"In the old days, in small data computing, you moved data to the computer," Nvidia Chief Executive Jensen Huang told Reuters. "But when you have giant amounts of data like Snowflake does, and the pile of proprietary data ... data that's so valuable to a company, then you move the compute to the data."
In this case, Nvidia is taking a "fairly engineering intensive" move of embedding its NeMo platform for training and running generative AI models into the Snowflake Data Cloud, said Huang.
The partnership comes as chatbot ChatGPT has pushed many companies to find their AI strategies and has propelled Nvidia, which provides the main hardware for AI, to becoming a trillion dollar company.
"This is significant. This is the last mile that we've been waiting for 40 years," said Frank Slootman, Chairman and CEO of Snowflake. "Every industry is on this. They used to say software is eating the world. Well, now data is eating software," he said about the importance of data today.
Slootman said companies that use Snowflake to manage their data will be able to now use their own data to train new AI models to gain an advantage in business without risking losing control of it.
No financial details of the partnership were disclosed, but Huang said Nvidia would benefit as more customers use computing for AI work.
"We sell more chips, and we have an operating system for AI called Nvidia AI Enterprise. And that operating system makes it possible for our chips to process AI," said Huang. Nvidia charges customers for the use of its Nvidia AI Enterprise software.
