NVIDIA unveiled the next generation of its upscaling technology, DLSS 5, at GTC 2026, marking a significant leap in AI-driven graphics. Dubbed '3D-guided neural rendering,' DLSS 5 shifts from traditional upscaling to fully AI-powered re-rendering, generating photorealistic lighting, materials, and shadows in real time from a fully rendered 2D frame and motion vectors. Unlike previous iterations, the technology focuses on enhancing existing visuals without fundamentally altering the rendering pipeline, using advanced neural networks to interpret and improve the image based on color data and motion cues.
The implementation of DLSS 5 currently demands substantial hardware: during the reveal, NVIDIA confirmed that two GeForce RTX 5090 graphics cards are required to run the technology. However, the company promises future optimizations will enable DLSS 5 to run on a single GPU. The technology will be available starting autumn 2026 on the latest GeForce RTX 5000 graphics cards and is backed by several major game developers, including Capcom, whose developers were both impressed and shocked by the DLSS 5 demo in Resident Evil Requiem, which dramatically altered the game's original artistic vision.
Despite its technical innovation, DLSS 5 has sparked intense debate. Critics, including the Art Director of Massive Damage Inc., argue that AI-enhanced rendering undermines developers' creative intent, leading to a uniform, overly realistic 'AI-slop' aesthetic. The Resident Evil Requiem demo exemplified this concern, with visuals so drastically changed that studios across the industry expressed alarm over increasing AI interference in game development. While some experts hail DLSS 5 as a milestone in computer graphics, others condemn it as AI-driven advertising that risks suppressing human creativity.
The internet has responded with a mix of amusement and criticism, with users humorously pointing out unintended visual quirks such as broad jawlines and hyperrealistic 2D styles. NVIDIA attempts to address these concerns by explaining that its AI operates solely on 2D image data and motion vectors, without direct access to 3D or depth information from the game engine. The company claims the AI understands the image well enough to apply improvements without altering the art design—a claim many developers and critics find unconvincing and contradictory. The controversy surrounding DLSS 5 extends beyond technical performance, raising ethical questions about AI's role in creative industries.