Nvidia and DataStax just made generative AI smarter and leaner. Here’s how

Nvidia And DataStax today launched a new technology that dramatically reduces storage requirements for companies deploying generative AI systems, while enabling faster and more accurate information retrieval in multiple languages.

The new one Nvidia NeMo Retriever microservicesintegrated with DataStax’s AI platformreduces data storage volume by 35 times compared to traditional approaches – a crucial capability as enterprise data is expected to become the more than 20 zettabytes in 2027.

“Current unstructured enterprise data is 11 zettabytes, roughly equivalent to 800,000 copies of the Library of Congress, and 83% of it is unstructured, with 50% of it being audio and video,” said Kari Briski, VP of product management for AI at Nvidia, in an interview with VentureBeat. “Significantly reducing these storage costs, while allowing companies to effectively embed and retrieve information, will be a game changer.”

The technology is already proving to be transformative Wikimedia Foundationthat used the integrated solution to reduce the processing time for 10 million Wikipedia articles from 30 days to less than three days. The system processes real-time updates from hundreds of thousands of entries edited daily by 24,000 volunteers worldwide.

“You can’t just rely on big language models for content; you need context from your existing business data,” explains Chet Kapoor, CEO of DataStax. “This is where our hybrid search capabilities come in, combining both semantic search and traditional text search, then using Nvidia’s re-ranker technology to deliver the most relevant results in real time on a global scale.”

Enterprise data security combined with AI accessibility

The partnership addresses a critical challenge facing companies: how to make their vast stores of private data accessible to AI systems without exposing sensitive information to external language models.

“Take FedEx: 60% of their data is in our products, including all package delivery information from the past 20 years with personal data. That won’t come to Gemini or OpenAI anytime soon or ever,” Kapoor explains.

The technology is quickly taking hold across sectors, with financial services firms leading the way despite regulatory limitations. “I’m impressed with how far the financial services industry has come,” says Kapoor Commonwealth Bank of Australia And Capital One as examples.

The next frontier for AI: multimodal document processing

Looking ahead, Nvidia plans to expand the technology’s capabilities to handle more complex document formats. “We’re seeing great results with multimodal PDF processing: we understand tables, graphs, charts and images and how they relate between pages,” Briski revealed. “It is a very difficult problem that we would like to tackle.”

For companies drowning in unstructured data as they try to deploy AI responsibly, the new offering provides a way to make their information assets AI-ready without compromising security or saving the bank too much on storage costs. The solution is immediately available via the Nvidia API Catalog with a 90-day free trial license.

The announcement underlines the growing focus on enterprise AI infrastructure as companies move beyond experimentation to large-scale deployment, with data management and cost efficiency becoming critical success factors.