Join top executives in San Francisco on July 11-12 to hear how leaders are integrating and optimizing AI investments for success. Learn more
Snowflake And Nvidia have partnered to provide companies with a platform to create custom generative artificial intelligence (AI) applications within the Snowflake Data Cloud using proprietary data. The announcement came out today Snowflake Summit 2023.
By integrating Nvidia’s NeMo platform for large language models (LLMs) and its GPU-accelerated computing with Snowflake’s capabilities, companies can use their data in Snowflake accounts to develop LLMs for advanced generative AI services such as chatbots, search, and summaries .
Manuvir Das, Nvidia’s head of enterprise computing, told VentureBeat that this partnership sets itself apart by enabling customers to customize their generative AI models via the cloud to meet their specific business needs. They can “work with their own data to… build industry-leading generative AI applications without taking them out of the secure Data Cloud environment. This reduces costs and latency while maintaining data security.”
Jensen Huang, founder and CEO of Nvidia, emphasized the importance of data in developing generative AI applications that understand each company’s unique operations and voice.
“Together, Nvidia and Snowflake will create an AI factory that helps companies turn their valuable data into custom generative AI models to power breakthrough new applications – directly from the cloud platform they use to run their businesses.” Huang said in a written statement.
>>Follow VentureBeat’s ongoing generative AI coverage<
According to Nvidia, the partnership offers enterprises new opportunities to leverage their own data, which can range from hundreds of terabytes to petabytes of raw and curated business information. They can use this data to create and refine custom LLMs, enabling business-specific applications and service development.
Streamline generative AI development through the cloud
Nvidia’s Das claims that enterprises using custom generative AI models trained on their own data will maintain a competitive advantage over companies that rely on vendor-specific models.
He said using fine-tuning or other techniques to customize LLMs creates a personalized AI model that allows applications to leverage institutional knowledge — the information gathered related to a company’s brand, voice, policies and operational interactions with customers. company.
“One way to think about modifying a model is to compare the output of a base model to a new employee who has just graduated from college, compared to an employee who has been with the company for more than 20 years” Das told VentureBeat. “The long-time collaborator has acquired the institutional knowledge necessary to solve problems quickly and with accurate insights.”
Creating an LLM involves training a predictive model using a huge corpus of data. Das said that to achieve optimal results, it’s essential to have abundant data, a robust model and accelerated computational capabilities. The new partnership encompasses all three factors.
“More than 8,000 Snowflake customers store exabytes of data in Snowflake Data Cloud. As companies look to add generative AI capabilities to their applications and services, this data fuels the creation of custom generative AI models,” said Das. “Nvidia NeMo runs on our accelerated computing platform and pre-trained base models will provide the software resources and computing power within Snowflake Data Cloud to make generative AI accessible to enterprises.”
Nvidia’s NeMo is a cloud-native enterprise platform that allows users to build, customize, and deploy generative AI models with billions of parameters. Snowflake plans to host and run NeMo within the Snowflake Data Cloud, enabling customers to develop and deploy custom LLMs for generative AI applications.
“Data is the fuel of AI,” says Das. “By building custom models using their data on Snowflake Data Cloud, enterprises can harness the transformative potential of generative AI to advance their business with AI-powered applications that deeply understand their business and the domains in which they operate.”
What’s next for Nvidia and Snowflake?
Nvidia also announced its commitment to offer accelerated computing and an expanded suite of AI software as part of the partnership. The company stated that substantial co-engineering efforts are underway, with the intention of integrating the Nvidia AI engine into Snowflake’s Data Cloud.
Das said generative AI is one of the most transformative technologies of our time, potentially impacting nearly every business function.
“Generative AI is a multi-trillion dollar opportunity and has the potential to transform any industry as enterprises begin to build and deploy custom models using their valuable data,” said Das. “As a platform company, we are currently helping our partners and customers harness the power of AI to solve humanity’s biggest problems with accelerated computing and full-stack software designed to address the unique needs of virtually every industry. “
VentureBeat’s mission is to become a digital city plaza where tech decision makers can learn about transformative business technology and execute transactions. Discover our Briefings.