Nvidia Pitches DLSS 5 Again, Promises to Preserve 'Artistic Intent' Nvidia at SIGGRAPH in Los Angeles detailed DLSS 5, its AI-powered graphics tool, promising to preserve 'artistic intent' after backlash at GTC earlier this year. The company demonstrated three model options and per-object controls, and claimed the one-step diffusion model can now run on a single GPU, addressing efficiency concerns. DLSS 5 is slated for a fall debut. Nvidia this week offered more details on DLSS 5, which uses AI to add photorealistic graphics to games. At the SIGGRAPH https://s2026.siggraph.org/ event in Los Angeles, Nvidia demonstrated DLSS 5's three models and how each uses a multi-step process to preserve the integrity of the game's visuals while allowing deep customization, VideoCardz reports https://videocardz.com/newz/nvidia-details-dlss-5-three-models-per-object-controls-and-single-gpu-support . DLSS 5 got off to a rough start at GTC earlier this year. Although we found it had potential in some settings /graphics-cards/164156/i-watched-nvidias-controversial-dlss-5-turn-potato-faces-into-photorealistic-characters , the wider gaming community didn't like how it seemed to strip games of artistic intent and prioritized an Instagram-filter-like look for characters, particularly female ones. At SIGGRAPH, Nvidia made artistic intent a key part of its DLSS 5 showcase. It highlighted how its new generative rendering tool uses a combination of the originally rendered frame and internal engine buffers to keep the generated result grounded in the original scene. It's also introduced a trio of model options for developers to choose from, each offering a slightly different feel and end result. Alongside this, Nvidia offers developers a number of sliders and toggles to adjust the intensity of the AI's output for each object and scene. Developers can also enable or disable DLSS 5 for specific elements within a scene. That means if a developer wanted to use DLSS 5 to improve certain assets but not others, they have that option. Nvidia says this works across traditionally rasterized and ray-traced games, too. Another big part of Nvidia's SIGGRAPH presentation was about making DLSS 5 more efficient to run. In Nvidia's original showcase, it needed the power of a second RTX 5090 just to demonstrate what many gamers ultimately considered AI slop. Nvidia now claims that DLSS 5's one-step diffusion model can run on a single GPU, thanks to it being restricted to improving the look of elements within a scene, rather than re-rendering the entire thing. It's possible the more advanced models will need more powerful hardware, but clearly, efficient operation is something Nvidia is working to improve. Regardless of what DLSS 5 can do, if it requires a second GPU, or even just one high-end GPU to run, its adoption will be vanishingly small for many years to come. The average gamer is still playing on RTX 3060 and 4060-equivalent cards. With current prices, most gamers aren't buying flagship GPUs. Nvidia has slated DLSS 5 for a fall debut. We'll have to see if its reception is any better the third time around.