Yesterday, on March 16, at the GDC 2026 conference, NVIDIA unveiled the next version of their image enhancement and scaling technology, Deep Learning Super Sampling (DLSS). They claim that DLSS 5 marks “the biggest breakthrough in computer graphics” since the advent of real-time ray tracing.
Example of DLSS 5 in Hogwarts Legacy
NVIDIA stated that DLSS 5 can render graphics photorealistic. The technology uses vector data from each frame of a game and applies a special generative artificial intelligence to enhance lighting and materials.
NVIDIA also demonstrated several examples of how DLSS 5 works in specific games like Resident Evil Requiem, Hogwarts Legacy, and Starfield. According to these demonstrations, the most noticeable changes are to character faces — they start to look more realistic, although this creates an “uncanny valley” effect.
The public’s reaction to DLSS 5 was mixed. Many online criticized NVIDIA, comparing the technology’s output to AI filters and calling the resulting images “neuroslop.” Additionally, some audience members expressed concerns that DLSS 5 might spoil the developers’ original intent. At the same time, there were defenders of DLSS 5. For instance, Todd Howard, head of Bethesda, expressed satisfaction with the results, stating he eagerly anticipates the tool’s release.
It has been confirmed that on launch, DLSS 5 will be supported in games by companies such as Bethesda, Capcom, Ubisoft, Warner Bros. Games, NetEase, NCSOFT, and Tencent.
The release of DLSS 5 is expected in fall 2026.