Nvidia unveils DLSS 5 to add AI-generated realism to games

In short: Nvidia announced DLSS 5, a new DLSS version that uses generative AI to create more realistic game visuals in real time.

What happened

Nvidia unveiled DLSS 5 at its GTC 2026 event. DLSS is Nvidia’s feature that helps games look better and run smoothly.

Earlier DLSS versions mainly worked by upscaling, which means taking a lower-resolution image and sharpening it to look like a higher-resolution one. Nvidia says DLSS 5 goes further by using generative AI, the same general idea as AI that can create images, to generate realistic details like lighting and surface texture.

Nvidia describes DLSS 5 as a “real-time neural rendering model.” In plain terms, it is an AI system that helps draw each frame of a game (like adding smarter paint strokes on top of a 3D scene). It uses structured graphics data from the game, including color and motion information for each frame, so the AI stays consistent with what is actually happening in the scene.

The company says this can improve hard-to-render details such as how light moves through skin, how fabric catches highlights, and how hair reflects light. DLSS 5 is designed to run during gameplay at up to 4K resolution.

Nvidia also says developers keep control, with settings for intensity, color look, and masks (rules that limit where the effect applies). DLSS 5 will use Nvidia’s existing Streamline integration system, and it is expected to arrive in fall 2026 with support from publishers like Bethesda, CAPCOM, Ubisoft, Tencent, and Warner Bros. Games.

Why it matters

If it works as promised, DLSS 5 could make games look more like real life without slowing them down as much. Nvidia CEO Jensen Huang also said similar methods could be used in film and design, which could help artists preview realistic scenes faster.

Source: TechCrunch AI