DLSS 5 Explained Clearly In 8 Minutes (How It Actually Works)

DLSS 5 is Nvidia’s new AI-powered graphics technology that goes beyond upscaling and frame generation by fundamentally reimagining game visuals for photorealism, but it has sparked controversy for altering character appearances and raising concerns about AI bias and artistic control. While it promises a major leap in visual fidelity and is supported by major studios, its resource demands and potential to override developers’ artistic intent have led to debate over who ultimately controls a game’s final look.

Nvidia unveiled DLSS 5 at GTC 2026, showcasing it with a demo of Resident Evil Aquam that immediately sparked controversy online. Viewers noticed that a character’s appearance had changed—her lips were fuller, cheekbones sharper, and makeup different—despite no update to the game itself. This was due to DLSS 5, a new AI-powered technology that Nvidia’s CEO Jensen Huang called the “GPT moment for graphics.” Unlike previous DLSS versions, which focused on performance improvements like upscaling and frame generation, DLSS 5 fundamentally alters the visual output of games, aiming for photorealism by reimagining lighting and materials.

DLSS 5 works by taking the frame rendered by the game engine, along with motion vectors, and passing them through a neural rendering model trained on Nvidia’s supercomputers. This model doesn’t just process pixels; it analyzes the entire scene, recognizing characters, fabrics, lighting conditions, and more. It then generates a new version of the frame with what it interprets as photorealistic lighting and material effects, such as subsurface scattering on skin and realistic fabric sheen—effects that are difficult to achieve even with advanced ray tracing due to computational demands.

The technology is currently resource-intensive, requiring two high-end RTX 5090 GPUs—one for rendering and one for running the neural model—making it accessible only to those with expensive hardware. Nvidia claims they will optimize DLSS 5 to run on a single GPU by launch, but for now, it’s a premium feature. The demo included games like Hogwarts Legacy, Starfield, and Oblivion Remastered, signaling broad industry support from major publishers.

Controversy erupted because DLSS 5 doesn’t just enhance visuals; it can alter the appearance of characters, raising concerns about AI bias and the imposition of certain beauty standards—dubbed the “Yasifi” effect. Critics worry that DLSS 5 breaks the promise of previous DLSS versions, which improved performance without changing artistic intent. Nvidia responded by emphasizing that developers retain full artistic control, with options for intensity sliders, color grading, and masking, and that DLSS 5 can be toggled off. However, there are concerns that modders could force DLSS 5 onto games without developer oversight.

DLSS 5 represents a significant shift toward neural rendering in the gaming industry, acting as a hybrid between traditional rendering and generative AI. While the technology is impressive and supported by major studios, it arrives amid widespread skepticism about AI-generated content in games. The debate now centers on who controls the final look of games: developers, players, or the AI itself. Despite the backlash, DLSS 5 is seen as a major leap forward in graphics, with the potential to redefine visual standards in gaming if its implementation and control are handled thoughtfully.