Nvidia
announced Monday that it is partnering with
Snowflake
to offer generative AI technology to the cloud-based data warehousing company’s customers.
In a recent videoconference call session with reporters, Manuvir Das,
Nvidia
‘s (ticker: NVDA) vice president of enterprise computing, spoke optimistically of the chipmaker’s new agreement with
Snowflake
(SNOW).
“Each end customer of Snowflake can create a model [suited] for themselves,” he said.
Snowflake is a leading vendor of data cloud and data analysis cloud services. It has more than 8,000 customers, as of April 2023.
Das said the partnership will enable Snowflake’s customers to securely build customized large language models using their own proprietary data. The deal is different from Nvidia’s partnership with
ServiceNow
(NOW), which was announced last month and offered a more standardized offering for IT management, customer service, and human resources applications, according to the executive.
Instead, each Snowflake model will be built, trained, and fine-tuned individually by each customer and be stored on Snowflake’s data cloud.
Snowflake will integrate Nvidia’s artificial intelligence software and chips to power the service. Nvidia didn’t specify when the new capabilities will be released.
Nvidia chips have high exposure to generative AI, which has been trending this year. The technology ingests text, images, and videos in a brute-force manner to create content.
Interest in this form of AI was sparked by OpenAI’s release of ChatGPT late last year. Chatbots like ChatGPT use a language model that generates humanlike responses, or their best guesses, based on word relationships found by digesting what has previously been written on the internet or from other forms of text.
Write to Tae Kim at [email protected]
Read the full article here