Nvidia unveiled a new version of its graphics technology called DLSS 5 during the company’s keynote presentation at the GTC conference. The system combines traditional 3D graphics data with generative AI models to produce detailed game visuals while reducing the computing resources required to render scenes.
Nvidia said the technology allows its GPUs to create realistic environments and characters by predicting portions of images rather than rendering every element directly.
Combining Structured Graphics Data And Generative AI
The DLSS system uses two different forms of data processing. Traditional rendering supplies structured 3D graphics information describing the virtual environment, while generative AI models generate additional visual details.
Jensen Huang said the combination of structured and probabilistic computing enables a different approach to rendering.
“We fused controllable 3D graphics, the ground truth of virtual worlds, the structured data … with generative AI, probabilistic computing,” Huang said during the keynote.
The system uses structured graphics data as a base layer and generative models to estimate and fill in parts of an image, which can reduce the computational workload while maintaining visual detail.
According to Huang, combining structured information with generative models allows developers to create content that remains controllable while still benefiting from AI-generated realism.
Gaming Origins And Broader Applications
Gaming historically played a central role in Nvidia’s development as a company, though the segment now represents a smaller portion of its revenue compared with newer markets such as artificial intelligence computing.
Huang presented DLSS 5 as an example of a broader shift in computing that could apply beyond video games.
The concept of merging structured data with generative AI models, he said, may be used across multiple industries.
“This concept of fusing structured information and generative AI will repeat itself in one industry after another,” Huang said.
Potential Role In Enterprise Data Systems
During the presentation, Huang also pointed to enterprise data platforms that rely heavily on structured information.
Examples included Snowflake, Databricks, and BigQuery.
Huang said future AI systems may combine structured enterprise databases with generative AI models to analyze information and produce insights.
“In the future, what’s going to happen is these data structures are going to be used by AI, and AI is going to be much, much faster than us,” Huang said.
He added that future AI agents could draw from both structured databases and generative datasets when performing analysis or completing tasks.
Featured image credits: Roboflow Universe
For more stories like it, click the +Follow button at the top of this page to follow us.