After the backlash around DLSS 5, Jensen Huang is now speaking much more carefully about the technology and claims he understands why players were upset. That does not mean NVIDIA is backing down: according to the company, this AI-driven direction is here to stay and will remain part of game development in the long term.
The reveal of DLSS 5 triggered a major backlash across the gaming community, mainly because many players felt that NVIDIA’s new technology was pushing far too deep into generative AI territory. The situation only got worse when Jensen Huang, the company’s founder and CEO, recently took a confrontational tone and argued that the public had completely misunderstood what he had previously “carefully explained”. Now, however, he has shifted to a noticeably different message.
The new comments came during Lex Fridman’s podcast, where Huang spoke in a far more understanding way about the controversy. He said that, in his view, the players’ perspective makes sense, and that he personally does not like AI “slop” either, meaning that increasingly same-looking, artificially inflated visual content that many people believed they were already seeing in DLSS 5. According to Huang, though, the misunderstanding comes from the idea that NVIDIA wants developers to build games the traditional way and then let the company’s system loose on top of them afterward.
NVIDIA Is Not Backing Down, It Is Just Selling It Differently
He argues that DLSS 5 is guided by 3D models, reference frames, textures, and the original artistic work, so the system is supposedly meant to stay faithful to what the developers created. Huang stressed that the geometry remains intact in every frame and that the goal of the technology is not to arbitrarily alter content, but to give creators another tool to work with. According to him, developers may eventually be able to train their own models and even request different visual styles, including something more cartoon-like, while still keeping the results consistent with the game’s intended art direction.
That still does not fully resolve the contradiction at the center of the debate. Huang may now sound much more careful and much less dismissive than he did before, but the core message has not changed: NVIDIA still wants generative AI to become part of game development in some form. The difference is mostly in how the company is now framing that ambition. Instead of presenting the technology as a kind of external post-process, NVIDIA is now trying to market it as a development tool that artists are supposedly meant to control.
It also became clear that NVIDIA does not imagine the future of DLSS 5 as being limited to photorealistic graphics. Huang says the system is open enough to support other styles as well, and that developers will be able to shape it around their own needs. Even so, the whole thing still feels fairly vague, especially because NVIDIA clearly does not want to define the boundaries too precisely before the technology actually arrives in autumn 2026. That is when it will become clear whether studios are really willing to trust this system inside their creative pipelines, and whether the end result feels like an actual improvement or just the source of another round of controversy.
Source: 3DJuegos
