🧔♂️ A friendly human may check it before it goes live. More news here
Nvidia unveils genAI system for real-time video game graphics
Nvidia unveiled DLSS 5 at its GTC keynote, a generative AI graphics system that mixes 3D scene data with predictive models to render detailed visuals with less compute.
The system combines traditional 3D graphics data with generative models that predict and fill parts of an image so GPUs need not render every element from scratch.
Jensen Huang, CEO, said the approach fuses structured 3D information with probabilistic generative AI to produce realistic yet controllable content.
Huang framed DLSS 5 as an example of a broader computing shift that could extend beyond gaming into enterprise use cases.
Huang pointed to enterprise data platforms such as Snowflake, Databricks, and BigQuery as examples that future AI agents could analyse and generate insights from, and said gaming is a smaller share of revenue than before.
🔗 Source: TechCrunch
🧠 Food for thought
Implications, context, and why it matters.
Nvidia’s lead in graphics upscaling sets up DLSS 5
- Before DLSS 5, Nvidia’s Deep Learning Super Sampling (DLSS) was widely seen as a top graphics upscaling option, using dedicated Tensor Cores (AI-focused processing units) on its RTX GPUs for AI-based rendering 1.
- AMD’s FidelityFX Super Resolution (FSR) competes with DLSS and stays hardware-agnostic, so it runs on a wider range of graphics cards 1.
- Head-to-head tests often find DLSS delivers cleaner image quality with fewer artifacts such as graining or flickering, especially at lower resolutions or performance-focused modes 2.
- In some cases, the DLSS AI-reconstructed image has looked better than the natively rendered version 1.
DLSS 5 links to Nvidia’s effort to run enterprise AI on proprietary data
- The DLSS 5 idea of using generative models on structured 3D data matches Nvidia’s enterprise work to help partners build and run AI on proprietary data (their internal, non-public datasets) inside existing platforms and workflows 3.
- Databricks (a data analytics platform) said it is adding native support for NVIDIA GPU acceleration on the Databricks Data Intelligence Platform, plus plans for native support for NVIDIA-accelerated computing in Photon (Databricks’ high-performance query engine) 4.
- Snowflake (a cloud data platform) and NVIDIA said they are working to help customers build customized generative AI applications using proprietary data in the Snowflake Data Cloud, including via NVIDIA NeMo (Nvidia’s toolkit for building and customizing generative AI models) and GPU-accelerated computing; Jensen Huang called it an “AI factory” 3.
- Nvidia lands as an infrastructure supplier for enterprise AI work, with partners wiring Nvidia hardware plus software into their generative AI products 3.
Stay updated on the go with our mobile app.
Get latest insights with smoother, more personalized experience through TIA mobile app.