Jensen Huang says developers keep artistic control
NVIDIA CEO Jensen Huang has responded to the early backlash around DLSS 5, dismissing criticism from gamers who say the technology changes how games and characters look. Speaking during a press Q&A with Tom’s Hardware at GTC 2026, Huang said concerns about DLSS 5 removing artistic intent are misplaced.
The criticism and floodgates for new DLSS memes appeared shortly after NVIDIA introduced DLSS 5 and showed how its neural rendering system can infer more photorealistic detail in games. Much of the reaction online focused on changes to character appearance in recent demos, including Resident Evil Requiem. Some, if not the majority of players, argued that the technology makes games look less attractive or pushes them toward the same visual style.
“Well, first of all, they’re completely wrong. The reason for that is because, as I have explained very carefully, DLSS 5 fuses controllability of the of geometry and textures and everything about the game with generative AI,”
— Jensen Huang (NVIDIA), during GTC 2026 Q&A Session
Huang said that interpretation is incorrect. According to him, DLSS 5 combines developer-authored geometry and textures with generative AI, while still leaving control in the hands of game creators . He said developers can fine-tune the model to match the intended art direction rather than hand that process over entirely to AI.
He also stressed that DLSS 5 is not a traditional post-processing effect applied after a frame is rendered. Instead, he described it as a geometry-level system with what NVIDIA calls content-controlled generative AI. Huang added that studios can experiment with different looks, including stylized or non-photorealistic outputs, depending on the type of game they want to build.
Gamers also have a choice. They can simply avoid enabling DLSS 5 in future games and rely on current upscaling techniques. Alternatively, they can choose other hardware that does not support this technology yet. It is also worth noting, though, that AMD recently said during the FSR Diamond announcement that its goal is to make everything ready for neural rendering. It’s not clear if they planned to make this part of the Diamond, though.
NVIDIA still has a lot to think about. The company still has time, and this is likely not the last discussion on the subject. I think the showcase would have landed better if NVIDIA had focused on non-human rendering first, rather than presenting clearly AI-generated faces in its first demos.
Source: Tom’s Hardware
