Sep 22, 2026 · 40 min
DLSS 5 tests the limits of AI-made game graphics
DLSS 5 explained, with Linus Tech Tips
Nvidia’s latest rendering push could improve visual fidelity while reshaping artistic control, hardware access, and the economics of game development.
- 1DLSS 5 moves beyond upscaling by generating or enhancing individual visual elements in games.
- 2Its facial changes expose a conflict between photorealistic enhancement and developers’ intended visual styles.
- 3Nvidia’s expanding feature ecosystem could deepen hardware dependence as developers and players chase higher fidelity.
Don't miss
Linus and David revisit DLSS 5’s controversial facial changes while arguing that future developer tuning could make the current version look primitive.
The brief
DLSS 5 extends Nvidia’s supersampling approach from reconstructing images to generating or enhancing individual game elements, prompting controversy over AI-made visual changes.
Linus Sebastian and David Gautier examine whether DLSS 5 preserves artistic intent, especially in stylized games that are not designed around photorealistic enhancement.
The technology’s hardware demands and Nvidia exclusivity raise practical questions about access, future GPU purchases, and whether optimization is being replaced by feature stacking.
The discussion connects DLSS 5 to rising development budgets, GTA 6, Expedition 33, and the growing cost of delivering high-fidelity graphics at scale.
The standout tension is simple: neural rendering may eventually make today’s facial changes look primitive, but buyers are warned not to purchase hardware for promises alone.
The episode leaves DLSS 5 as both a rendering breakthrough and an ecosystem strategy, with its value depending on developer tuning and broader competition.