Nvidia introduced DLSS 5 at GTC 2026 as its biggest graphics advancement since ray tracing. In a matter of hourstechnology became the biggest meme generator of the year and turned the entire gaming community against it. After the criticism, the CEO of Nvidia has come to justify the company’s decisions. Spoiler: the situation has not improved.
Almost unanimous rejection. Nvidia presented DLSS 5 at GTC 2026 as an evolution that goes beyond upscaling traditional. The system analyzes the color and motion vectors of each frame and applies an AI model to generate photorealistic lighting and materials anchored to the game’s original 3D data. However, a good part of the community and professionals industry called DLSS 5 “AI slop.”
The examples used by Nvidia as a demonstration have been the spearhead of these problems: Grace Ashcroft from ‘Resident Evil Requiem’ with more pronounced cheekbones, fuller lips and uniform skin, or characters from ‘Starfield’ who gain facial resolution but at the cost of aesthetic coherence… A fundamental error that came from assuming that “more photorealistic” is equivalent to “better”, when decades of video game design point in a direction that is, at the very least, complementary.
Nvidia’s response. Jensen Huang responded to the controversy during a question session in GTC 2026 itself, in what was the CEO’s first public reaction to the general unrest. His words left no room for ambiguity: “Well, first of all, people are completely wrong.” Their technical argument was that, according to Huang, DLSS 5 combines developer-created geometry and textures with generative AI, leaving control in the hands of art teams. Developers can adjust the model to fit the intended art direction, rather than handing that process over entirely to AI.
The manager insisted that It’s not about post-processing at the frame level.but rather generative control at the geometry level, what Nvidia calls “content-controlled generative AI”, adding generative capacity over the game’s existing geometry without altering artistic control. Huang added that studios can also explore stylized or non-photorealistic results depending on the type of game they want to build, and that players who don’t want to activate DLSS 5 they can just ignore it and continue with current scaling techniques. Bethesda supported that position. The study clarified that what was shown was a very preliminary version.
Calculation errors. There were those who pointed out that the presentation of DLSS 5 in the middle of a keynote mostly dedicated to the cloud and enterprise AI, before an audience already extremely polarized with AI, turns Huang’s statements into ammunition that will be repeated until Nvidia manages to convincingly explain what DLSS 5 is and why gamers and developers need it. Comparisons have been read to Apple’s “Antennagate” (when the company’s response to criticism was perceived as arrogance): being right and communicating it poorly is like not being right.
What’s behind. PC Gamer also points to a deeper debate: if it was previously possible to modify the graphical style of a game through adjustments or mods, hardware-assisted generative AI represents a shortcut controlled by large technology companies that could radically transform how artistic authorship is perceived in video games. It remains to be resolved whether when games arrive that have integrated DLSS 5 from the beginning and with a defined artistic direction they will change that perception. There it will be detected, without a doubt, who in this entire battle was really wrong.
In Xataka | NVIDIA is going to spend $4 billion on photonics companies. He is preparing for what is coming
GIPHY App Key not set. Please check settings