Monday, March 16th 2026

NVIDIA Unveils DLSS 5 with Real-Time Neural Rendering

NVIDIA today unveiled NVIDIA DLSS 5, the company's most significant breakthrough in computer graphics since the debut of real-time ray tracing in 2018. DLSS 5 introduces a real-time neural rendering model that infuses pixels with photoreal lighting and materials. Bridging the divide between rendering and reality, DLSS 5 empowers game developers to deliver a new level of photoreal computer graphics previously only achieved in Hollywood visual effects.

"Twenty-five years after NVIDIA invented the programmable shader, we are reinventing computer graphics once again," said Jensen Huang, founder and CEO of NVIDIA. "DLSS 5 is the GPT moment for graphics—blending hand-crafted rendering with generative AI to deliver a dramatic leap in visual realism while preserving the control artists need for creative expression."
Bridging The Cinematic Gap
Since the dawn of GeForce, NVIDIA has strived to deliver the graphics horsepower required for game developers to create incredible, realistic worlds—where lighting, reflections and shadows obey the laws of nature.

From programmable shaders with GeForce 3 in 2001, to CUDA with GeForce 8800 GTX in 2006, to real-time ray tracing with GeForce RTX 2080 Ti in 2018, to path tracing and neural shaders with GeForce RTX 5090 in 2025, NVIDIA has delivered major architectural innovations and a massive 375,000x increase in compute to meet this challenge.

However, the rendering horsepower available to a 16-millisecond game frame remains a tiny fraction of that available to a photoreal Hollywood VFX frame, which can take minutes to hours to render. Real-time rendering cannot bridge the gap to photorealism through brute force alone.

NVIDIA DLSS was released in 2018 as an AI technology to boost performance, first by upscaling resolution, and then by generating entirely new frames. It has been integrated in over 750 games, becoming a gold standard for the industry. Launched at CES this year, DLSS 4.5 uses AI to draw 23 out of every 24 pixels seen on the screen. Today, DLSS is evolving beyond performance to transform visual fidelity in games.

Video AI models have rapidly learned to generate photoreal pixels, but they run offline, are difficult to precisely control and often lack predictability, with every new prompt generating bespoke content. For games, pixels must be deterministic, delivered in real time and tightly grounded in the game developer's 3D world and artistic intent.

DLSS 5 takes a game's color and motion vectors for each frame as input, and uses an AI model to infuse the scene with photoreal lighting and materials that are anchored to source 3D content and consistent from frame to frame. DLSS 5 runs in real time at up to 4K resolution for smooth, interactive gameplay.

The AI model is trained end to end to understand complex scene semantics such as characters, hair, fabric and translucent skin, along with environmental lighting conditions like front-lit, back-lit or overcast—all by analyzing a single frame. DLSS 5 then uses its deep understanding to generate visually precise images that handle complex elements such as subsurface scattering on skin, the delicate sheen of fabric and light-material interactions on hair, all while retaining the structure and semantics of the original scene.

DLSS 5 provides game developers with detailed controls for intensity, color grading and masking, so artists can determine where and how enhancements are applied to maintain each game's unique aesthetic. Integration is seamless, using the same NVIDIA Streamline framework used by existing DLSS and NVIDIA Reflex technologies.

Availability & Game Developer Support
DLSS 5 will be supported by the industry's biggest publishers and game developers, including Bethesda, CAPCOM, Hotta Studio, NetEase, NCSOFT, S-GAME, Tencent, Ubisoft and Warner Bros. Games.

"Bethesda has such a rich history pushing graphics with NVIDIA, going all the way back to Morrowind, with that incredible water," said Todd Howard, studio head and executive producer at Bethesda Game Studios. "When NVIDIA showed us DLSS 5 and we got it running in Starfield, it was amazing how it brought it to life. We've played it. We can't wait for all of you to do so as well."

"At CAPCOM, we strive to create experiences that feel cinematic, compelling and deeply believable—where every shadow, texture and ray of light is crafted with intention to enhance atmosphere and emotional impact," said Jun Takeuchi, executive producer and executive corporate officer at CAPCOM. "DLSS 5 represents another important step in pushing visual fidelity forward, helping players become even more immersed in the world of Resident Evil."

"Immersion is about making the world feel real. DLSS 5 is a real step towards that goal," said Charlie Guillemot, co-CEO of Vantage Studios. "The way it renders lighting, materials and characters changes what we can promise to players. On Assassin's Creed Shadows, it's letting us build the kind of worlds we've always wanted to."

DLSS 5 will come to games including AION 2, Assassin's Creed Shadows, Black State, CINDER CITY, Delta Force, Hogwarts Legacy, Justice, NARAKA: BLADEPOINT, NTE: Neverness to Everness, Phantom Blade Zero, Resident Evil Requiem, Sea of Remnants, Starfield, The Elder Scrolls IV: Oblivion Remastered, Where Winds Meet, and more.

DLSS 5 will arrive this Fall. See the first preview this week at GTC, and below, you can check out several examples of NVIDIA DLSS 5 in Resident Evil Requiem, EA SPORTS FC, Starfield, Hogwarts Legacy, and the NVIDIA Zorah tech demo.

Resident Evil Requiem
EA SPORTS FC
Starfield
Hogwarts Legacy
NVIDIA Zorah Tech Demo
Source: NVIDIA
Add your own comment

305 Comments on NVIDIA Unveils DLSS 5 with Real-Time Neural Rendering

#251
Visible Noise
RGAFLI'm also 'In The Know' so to speak and the way games are being made is not changing for the forseeable future. If raster is tapped out why release a PS6.
You mean the PS6 that’s going to focus on pathtracing?

I’ll give you a moment to Google who Mark Cerny is.

You didn’t answer my question, although I’ll admit I didn’t expect one.
Posted on Reply
#252
LastDudeALive
theglazeDude, you're comparing average FPS between 2x Frame Generation (4090) to 4x Frame Generation (5070).
Yes! Exactly! The 5070 has new features the 4090 does not. Those features enable it to reach higher frame rates than the 4090. That's exactly what the 5070=4090 slide at CES was saying!
Posted on Reply
#253
RGAFL
Visible NoiseYou mean the PS6 that’s going to focus on pathtracing?

I’ll give you a moment to Google who Mark Cerny is.

You didn’t answer my question, although I’ll admit I didn’t expect one.
How can you focus on Path Tracing when you need the geometry, textures and material maps so that the light reacts. Please, go and do some research on how games are made and then come back. Path Tracing is how light reacts to a surface and how it's scattered at source. Those surfaces are textures with material maps applied. Guess what those are. I'll give you a clue and it begins with R.
Posted on Reply
#254
Visible Noise
RGAFLHow can you focus on Path Tracing when you need the geometry, textures and material maps so that the light reacts. Please, go and do some research on how games are made and then come back. Path Tracing is how light reacts to a surface and how it's scattered at source. Those surfaces are textures with material maps applied. Guess what those are. I'll give you a clue and it beigins R.
Ask the guy who is in charge of creating the PS6 - his name is Mark Cerny. He’s the person that designed the PS2 graphics engine, and is the lead designer of the PS4, PS5, and PS6.

Would you like to share your credits?

If you look really hard (not really) you’ll find an hour long presentation he gave on the design of the PS6 where he explicitly says rasterized graphics are tapped out.

As I said previously I think he knows a wee bit more about where game graphics are headed than some anonymous internet poster that doesn’t seem to know that materials don’t require rasterization when pathtracing, geometry isn’t limited to rasterization, nor have an understanding of neural rendering.

Goodbye.
Posted on Reply
#255
tpa-pr
As with seemingly every AI product/technology that gets released, this seems like a solution seeking a problem. As someone said earlier in the thread (or another thread, there are a lot of threads) we've had a solution for photorealistic face texturing for years now.

I'm just gonna stick to my rasterized games with textures. If a game requires DLSS 5 it'll be a no-buy from me. The backlog is large enough to cover my entire lifetime.
Posted on Reply
#256
RGAFL
Visible NoiseAsk the guy who is in charge of creating the PS6 - his name is Mark Cerny. He’s the person that designed the PS2 graphics engine, and is the lead designer of the PS4, PS5, and PS6.

Would you like to share your credits?

If you look really hard (not really) you’ll find an hour long presentation he gave on the design of the PS6 where he explicitly says rasterized graphics are tapped out.

As I said previously I think he knows a wee bit more about where game graphics are headed than some anonymous internet poster that doesn’t seem to know that materials don’t require rasterization when pathtracing, nor understand neural rendering.

Goodbye.
If you don't take my word for it then why not believe Google.

Yes, in the context of 3D computer graphics and rendering, materials (textures, colors, shaders) generally require rasterization to be displayed on a screen.

Rasterization is the fundamental process of converting 3D models—which include the materials and textures applied to them—into a two-dimensional, pixel-based image (raster format) for display.

Here is a breakdown of why this is necessary:
  • Real-time Rendering: In gaming and real-time engines, rasterization is the primary technique used to map materials (textures, colors, lighting) onto 3D geometry and convert them into pixels quickly.
  • Screen Display Requirement: Digital screens only support raster graphics (pixels). Therefore, vector-based materials or 3D models must be converted into a fixed pixel pattern (rasterized) for output.
  • Shading & Textures: Pixel shaders calculate the final color of each pixel on a 3D object, accounting for lighting, texture mapping, and physical properties, which are then rasterized into pixels.
  • Performance: Rasterization is favored for real-time applications because it is efficient, though alternative methods like ray tracing are increasingly used for improved visual quality.
I'm very aware of who Mark Cerny is thank you, bumped into him at Sony presentations once or twice. Do you?

I'll let the guys at Epic and Unity know that someone here has a game engine that can make games with pure Path Tracing. They might be in touch after laughing themselves to death.
Posted on Reply
#257
LastDudeALive
RGAFLIf you don't take my word for it then why not believe Google.

Yes, in the context of 3D computer graphics and rendering, materials (textures, colors, shaders) generally require rasterization to be displayed on a screen.

Rasterization is the fundamental process of converting 3D models—which include the materials and textures applied to them—into a two-dimensional, pixel-based image (raster format) for display.

Here is a breakdown of why this is necessary:
  • Real-time Rendering: In gaming and real-time engines, rasterization is the primary technique used to map materials (textures, colors, lighting) onto 3D geometry and convert them into pixels quickly.
  • Screen Display Requirement: Digital screens only support raster graphics (pixels). Therefore, vector-based materials or 3D models must be converted into a fixed pixel pattern (rasterized) for output.
  • Shading & Textures: Pixel shaders calculate the final color of each pixel on a 3D object, accounting for lighting, texture mapping, and physical properties, which are then rasterized into pixels.
  • Performance: Rasterization is favored for real-time applications because it is efficient, though alternative methods like ray tracing are increasingly used for improved visual quality.
I'm very aware of who Mark Cerny is thank you, bumped into him at Sony presentations once or twice. Do you?

I'll let the guys at Epic and Unity know that someone here has a game engine that can make games with pure Path Tracing. They might be in touch after laughing themselves to death.
And we've hit the limit on how good real-time rasterization can get for this entire pipeline. DLSS 5 involves replacing most of that chain with RT and AI. Only the basic 3D geometry is modeled and then rasterized. The rest of the scene is handled by PT and AI. A few PT rays provide a guide for how the scene is lit, and the AI creates the textures and models the lighting and material properties.
Posted on Reply
#258
RGAFL
Lighting is RT,PT or raster. Basically, the rest of the scene (geometry, textures, material maps) is still raster whether it's AI generated or not. The reason Path Tracing is taking over for lighting is because it's the most realistic way of doing lighting but also the most computationally expensive.

I'll tell you what, go and make me a pure Path Traced or Ray Traced game. No geometry, textures or material maps. Use whatever engine you like. If your so sure that it can be done surely you must know of a pure RT/PT game engine. I'll bet you £100 to a charity of your choice it won't be done.

Maybe Mark Cerny could help with that, I doubt it, but you could ask.
Posted on Reply
#259
Visible Noise
RGAFLI'll let the guys at Epic and Unity know that someone here has a game engine that can make games with pure Path Tracing.
Cool, I’ll let Andrew Lauritzen know you’ll be showing up on the dev forum. It’ll be fun watching him ban you after three posts.

Plonk.
Posted on Reply
#260
RGAFL
Visible NoiseCool, I’ll let Andrew Lauritzen you’ll be showing up on the forum. It’ll be fun watching him ban you after three posts.

Plonk.
So you lose an argument and get someone banned. Good one. Take it like a man.
Posted on Reply
#261
MufasaV
RGAFLSo why do Nvidia, AMD and Intel keep adding more raster cores as well as Tensor cores if a 3090/6900XT is in your view the pinnacle of raster. The 5090 is double the core count of the 3090, the 5080 has 23% more than the 3080 and roughly 10% more than the 4080. Think they know something you don't.
The only "raster cores" in a GPU are ROPS. Of which there have been the same number per SM or Compute Unit for over 10 years.

Math Units are general purpose and very useful for AI and RT as well as anything else.

Easter does not refer to the entire compute pipeline. Only the final transformation of a scene to fit a flat monitor
Posted on Reply
#262
Dr. Dro
Whoa, this thread took a turn lol. I'll provide my thoughts, nonetheless.

Even though pure raster graphics are tapped out (i'll die on this hill, sorry), it doesn't necessarily mean that gamers suddenly have advanced RT hardware at their disposal, nor the resources to obtain it either. It's a long part of the reason why people are so bothered by ray tracing, it is extremely costly (in terms of hardware power and in turn, $$$), current implementations just cannot shine due to hardware limitations, further slowing adoption. Economy isn't doing great either so disposable income is very limited to throw at things that, to the perception of most gamers, they can really live without.

It stands to reason, then, that it's not yet commercially viable for most publishers to ship games that require RT capable hardware. Personally, I can only think of two: Metro Exodus' enhanced edition (which is complimentary to the original) and Indiana Jones and the Great Circle - first of which is a DirectX 12 Ultimate title, the latter runs on Vulkan (requires VK_KHR_ray_tracing_pipeline, VK_KHR_ray_query and VK_KHR_dynamic_rendering, which are unsupported by both AMD's and Nvidia's Windows driver on GPU architectures earlier than Turing/RDNA 2 respectively).

Add RT to GenAI, and Nvidia's "up to 1 million times faster than Pascal" mythological architecture may really be necessary to truly master RT and generative graphics. Can they pull it off? I don't know. But I do know is that if any company can, it's Nvidia. If they don't pull that off... it'll simply never happen.
Posted on Reply
#263
GoldenTiger
Dr. Drogames that require RT capable hardware. Personally, I can only think of two: Metro Exodus' enhanced edition (which is complimentary to the original) and Indiana Jones and the Great Circle
There's also avatar and star wars outlaws :).
Posted on Reply
#264
Visible Noise
Dr. DroPersonally, I can only think of two: Metro Exodus' enhanced edition (which is complimentary to the original) and Indiana Jones and the Great Circle
Add the Final Fantasy remake.
Posted on Reply
#265
medi01
How could someone release this, then claim something about "realistic lighting" is beyond me.

Posted on Reply
#266
meem55
Well, the pointing to each other (Nvidia <> game makers) has started. According to this article, developers learned about the whole DSLL 5 the same time we did.
Quote from article:
Speaking with developers and artists at studios that have agreed to DLSS 5, including CAPCOM and Ubisoft, Insider Gaming was told that the DLSS 5 tech was revealed to them at the same time as everyone else.

“We found out at the same time as the public,” said one Ubisoft developer.

Developers at CAPCOM tell Insider Gaming that the announcement and the publisher’s involvement were particularly shocking, as CAPCOM has previously been historically very “anti-AI” with projects such as Resident Evil Requiem and other unannounced projects in development. Some at the publisher fear that the DLSS 5 announcement could prompt a change in the publisher’s view on generative AI and its implementation in its games.
Posted on Reply
#267
AVATARAT
As we’ve seen, the “nice” filter in DLSS 5 doesn’t care about the light source, so whether it’s better than the original is also a matter of subjective opinion.
Furthermore, the main idea behind DLSS is to achieve a higher FPS, not to lower it.
Nvidia simply made a wrong decision to include this filter in DLSS instead of just creating another software filter.
Posted on Reply
#268
lambda
Guwapo77...that wouldn't be fault of DLSS though.
Indeed, all I meant to say was it won't enhance immersion the way you expect it to. A game is not just about photorealistic graphics rather the whole experience.
Posted on Reply
#269
R0H1T
AVATARATAs we’ve seen, the “nice” filter in DLSS 5 doesn’t care about the light source, so whether it’s better than the original is also a matter of subjective opinion.
Furthermore, the main idea behind DLSS is to achieve a higher FPS, not to lower it.
Nvidia simply made a wrong decision to include this filter in DLSS instead of just creating another software filter.
The point's actually (probably?) to make Nvidia another key figure in games' development, beyond where they already are! Of course this will hurt AMD as well as Intel, Apple, QC et al.
Posted on Reply
#270
Dr. Dro
AVATARATAs we’ve seen, the “nice” filter in DLSS 5 doesn’t care about the light source, so whether it’s better than the original is also a matter of subjective opinion.
Furthermore, the main idea behind DLSS is to achieve a higher FPS, not to lower it.
Nvidia simply made a wrong decision to include this filter in DLSS instead of just creating another software filter.
"DLSS 5" is not a filter, it is not overlaid on top of the graphics, nor is it a replacer... it's geometry-level generative, and requires full developer intervention to be implemented. Fine grained controls to enhancements are provided. Through it, GenAI is the brush, not the artist.

By "DLSS" you're referring to the upscaler component, but it is an umbrella term (just like FidelityFX or XeSS) and has been for some time, we have:

DLSS Super Resolution
DLSS Ray Reconstruction
DLSS Frame Generation

All of which are independent of one another. DLSS Super Resolution is the original and what gets the most attention, with the latest CNN model being referred to in marketing material as "DLSS 3.5", the first transformer model as "DLSS 4" and the improved Gen 2 transformer as "DLSS 4.5", the implementation however is pretty much the same, you just need to have the updated DLL placed and use the correct render preset.

Moreover, it's clearly in development, let's reserve our judgment for when it releases. I promise I will be extra harsh if it gives games "stable diffusion syndrome".
Posted on Reply
#271
AVATARAT
Dr. Dro"DLSS 5" is not a filter, it is not overlaid on top of the graphics, nor is it a replacer... it's geometry-level generative, and requires full developer intervention to be implemented. Fine grained controls to enhancements are provided. Through it, GenAI is the brush, not the artist.

By "DLSS" you're referring to the upscaler component, but it is an umbrella term (just like FidelityFX or XeSS) and has been for some time, we have:

DLSS Super Resolution
DLSS Ray Reconstruction
DLSS Frame Generation

All of which are independent of one another. DLSS Super Resolution is the original and what gets the most attention, with the latest CNN model being referred to in marketing material as "DLSS 3.5", the first transformer model as "DLSS 4" and the improved Gen 2 transformer as "DLSS 4.5", the implementation however is pretty much the same, you just need to have the updated DLL placed and use the correct render preset.

Moreover, it's clearly in development, let's reserve our judgment for when it releases. I promise I will be extra harsh if it gives games "stable diffusion syndrome".
You can try to defend this geometry-bla-bla AI slop by Nvidia, but that won’t change the end result, which is complete rubbish. Besides, you can’t “promise” anything you haven’t been involved in developing :)
Posted on Reply
#272
Dr. Dro
AVATARATYou can try to defend this geometry-bla-bla AI slop by Nvidia, but that won’t change the end result, which is complete rubbish. Besides, you can’t “promise” anything you haven’t been involved in developing :)
You can dislike it all you want, just don't spread/reinforce misinformation tbh. It can and will be better, we're only judging it this harshly at a beta stage because it is Nvidia.
Posted on Reply
#273
R0H1T
Dr. DroThrough it, GenAI is the brush, not the artist.
Well technically Nvidia (JHH?) is the artist, especially with what they fed their abominations with!
Posted on Reply
#274
Dr. Dro
R0H1TWell technically Nvidia (JHH?) is the artist, especially with what they fed their abominations with!
Sigh, this is so frustrating. It's just mass hysteria at this point
Posted on Reply
#275
R0H1T
"AI skepticism" isn't really hysteria & I can promise you would be among those who're on the wrong side of history with this one!
Posted on Reply
Add your own comment
Nov 20th, 2026 15:59 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts