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

#76
BoxCutter
Finally, the real time Balensiaga meme.
Posted on Reply
#77
sbacc
Visible NoiseImagine that, changing lighting - such as self shadowing - could make a face look different!
Nice try, but actually knowing how light change a face is my work expertise, and it isn't the case here, but you would know if you had actually looked at the timestamp example in the linked video...

You could have just said that even if the face was changed, you still like much better the DLSS 5 one...
Posted on Reply
#78
Testsubject01
evernessince

The AI is pretty clearly changing the content (which doesn't make a lot of sense if they are only feeding it lighting information). Same issue stable diffusion runs into.

The lighting all has that AI look to it as well. I generate a ton of images locally and this to me looks like those lazy images where people are using the default art style.

I predicted lighting would be the next thing Nvidia would AI generate but I expected better result than this. Watching the video of this just makes me feel queezy (might be frame gen contributing as well to that).
Changing the creators visual intent is usually a big NONO in the professional industry anyway.

Especially since gen Ai will show a different result between each instance a model is shown in the game. (most likely why longer cutscenes/ gameplay was omitted in this showpiece as well).
Then showing FIFA, which even has expensive licences on the likeness of the featured players and advertises that a large part… good job slopping faces with Ai here.
Posted on Reply
#79
lambda
Guwapo77I see this different than most, I think the characters look great with DLSS 5 On. The problem I have is the environment...everything brighter and it feels like it removes the atmospheric effects.
They'll still have wooden stuff animation. It'll make the whole thing immersion breaking since the characters might look more "real" but they don't act like it. Combine it with shoddy voice work and we are primed for generic AI Slop being unleashed en masse.

I'm quite excited for this still, but I hope it's not used as a crutch for shoddy asset work.
Posted on Reply
#80
Youseethis
Visible NoiseDevs are ruining their own creative vision?

How exactly does that work??
When you replace the “vision” with pre-trained ai you are left with the machine’s constrct. Look at the Leon pictures as an example, from brooding to reallistic and hopeful.
Posted on Reply
#81
rv8000
This is a particularly embarrassing example on the lighting

DLSS 5 example seems to have no direct light source and shadowing is ruined. If the original image has the source of light shining on the left most face of the building, the DLSS 5 example has both faces lit and shadowed nearly the same.
Posted on Reply
#82
lambda
LastDudeALiveOn the bright side, this may stop developers from pushing even more polygons into their games in the quest for more "realism." Instead of relying on 1 trillion polygons to have micron-accurate 3D models and Lumen tanking performance on a 5090, they can have less detail and let DLSS do the work of making it more accurate.
The performance requirements would be similar if not higher for real-time image generation.
Posted on Reply
#83
Visible Noise
YouseethisWhen you replace the “vision” with pre-trained ai you are left with the machine’s constrct. Look at the Leon pictures as an example, from brooding to reallistic and hopeful.
You understand this is implemented by the developers, right? Are you paying attention?
Posted on Reply
#85
Visible Noise
lambdaThe performance requirements would be similar if not higher for real-time image generation.
Why? This isn’t image generation. This is correct lighting based upon ground truth training.

Think of it as Ray Reconstruction 2.0. Or in five years when AMD has it, Ray Regeneration 4.0.
Posted on Reply
#87
BoxCutter
Also, stills only, I wonder why.
Edit. NVM there is a video and it's about as good as one would expect.
Posted on Reply
#88
lambda
Visible NoiseWhy? This isn’t image generation. This is correct
lighting based upon ground truth training.
I'm unsure, it looks like AI assisted infill operation to me.
Posted on Reply
#89
Visible Noise
lambdaI'm unsure, it looks like AI assisted infill operation to me.
Look closer. Most of the difference is self shadowing and sub-surface light scattering.
Posted on Reply
#90
Youseethis
Visible NoiseYou understand this is implemented by the developers, right? Are you paying attention?
Then they created a schizophrenic contradiction within their work. Or is their vision to rewrite the story by “changing visual fidelity” ? Whoever owns a 5090 gets the true ending.
:toast:
Posted on Reply
#91
Visible Noise
Man, the amount of knee jerk, it’s different so it’s bad, responses are amazing.

Give it some thought people.
Posted on Reply
#92
rv8000
Visible NoiseWhy? This isn’t image generation. This is correct lighting based upon ground truth training.

Think of it as Ray Reconstruction 2.0. Or in five years when AMD has it, Ray Regeneration 4.0.
Considering the lighting is wrong in their own tech demo in the courtyard scene (quite blatantly), and also seems very off in the FIFA examples (direct source of light and direction seem completely gone), there is nothing correct about the lighting.
Posted on Reply
#93
Visible Noise
YouseethisThen they created a schizophrenic contradiction within their work. Or is their vision to rewrite the story by “changing visual fidelity” ? Whoever owns a 5090 gets the true ending.
:toast:
Wut?

The devs are now wrong for creating the look they want?

Ok, you do you, I guess.
rv8000Considering the lighting is wrong in their own tech demo in the courtyard scene (quite blatantly), and also seems very off in the FIFA examples (direct source of light and direction seem completely gone), there is nothing correct about the lighting.
The devs have their own lighting wrong. Got it.
Posted on Reply
#94
BoxCutter
Visible NoiseMan, the amount of knee jerk, it’s different so it’s bad, responses are amazing.

Give it some thought people.
It's bad, so it's bad. It's not rocket science. Stop baiting.
Posted on Reply
#95
esserpain
Haven’t been on this site in awhile, but I’ve decided to comment this time, because I have no end of concerns when it comes to how DLSS 5 is being promoted, and how it’s already slated to be used in games.

I’m not solely talking about how the example pictures look, BTW, despite that being a big reason why I’m concerned about this (see Grace’s face for an example of how bad DLSS 5 looks in its current state). Mainly, I’m referring to which developers are adopting it and how. Developers that make very stylized games are already jumping to use a technology that by all indications, will only really be at all useful and effective in photorealistic titles. Those developers and examples of games with stylized or semi-stylized visuals are underlined below.




The main issue I have with this is the potential for DLSS 5 to distort these stylized games - with very specific artstyles predetermined by the developers - into the same, ChatGPT Ghibli-ified sludge. I know that NVIDIA is claiming that developers will have the choice to use the models provided to them how they’d like (if I read the article correctly, that is - feel free to correct me if I’m wrong). But giving developers the tools to be lazy and to compromise their own creative potential for maximum profits and efficiency, and promoting the laziest developers on your public platforms as being the best NVIDIA partners, feels like a recipe for disaster.

I might be proven wrong, but judging by how previous DLSS versions became widely adopted and accepted across the industry and have become cost-cutting tools to let developers not put as much effort into optimization… I don’t expect to be.
Posted on Reply
#96
sbacc
esserpainHaven’t been on this site in awhile, but I’ve decided to comment this time, because I have no end of concerns when it comes to how DLSS 5 is being promoted, and how it’s already slated to be used in games.

I’m not solely talking about how the example pictures look, BTW, despite that being a big reason why I’m concerned about this (see Grace’s face for an example of how bad DLSS 5 looks in its current state). Mainly, I’m referring to which developers are adopting it and how. Developers that make very stylized games are already jumping to use a technology that by all indications, will only really be at all useful and effective in photorealistic titles. Those developers and examples of games with stylized or semi-stylized visuals are underlined below.




The main issue I have with this is the potential for DLSS 5 to distort these stylized games - with very specific artstyles predetermined by the developers - into the same, ChatGPT Ghibli-ified sludge. I know that NVIDIA is claiming that developers will have the choice to use the models provided to them how they’d like (if I read the article correctly, that is - feel free to correct me if I’m wrong). But giving developers the tools to be lazy and to compromise their own creative potential for maximum profits and efficiency, and promoting the laziest developers on your public platforms as being the best NVIDIA partners, feels like a recipe for disaster.

I might be proven wrong, but judging by how previous DLSS versions became widely adopted and accepted across the industry and have become cost-cutting tools to let developers not put as much effort into optimization… I don’t expect to be.
If next-gen console are much better at RT or "PT", Nvidia will need a new "thing" to differentiate their product enough from console, beyond the usual, a bit more resolution, higher FPS beyond 60 or 120 and slightly better RT effect.
This is the goal of that, I will be incorporated in probably a lot of new game in the future, but because of console it will never be mandatory anytime soon. So if anyone don't like the effect, they can just turn it off even on their "Nvidia PC". Turning it off by default as a mindset, will help save a few bucks on GPU purchase...
Posted on Reply
#97
Sithaer
YouseethisThen they created a schizophrenic contradiction within their work. Or is their vision to rewrite the story by “changing visual fidelity” ? Whoever owns a 5090 gets the true ending.
:toast:
In the Digital Foundry video I've watched they directly mentioned Todd Howard that he does in fact likes this DLSS 5 version of Starfield better than the original and he would prefer to play it like this..
Personally I'm curious to see how the final version will work cause its just too early to judge it in my eyes, in some games it could be actually better imo if done right. 'Not that my 5070 will run this but thats another story :laugh: '

Btw I don't really see how this is much more different than say using GFX/Texture/model and whatever else mods for a game.
I've always played my Elder Scrolls games with a bunch of mods and to me it was better that way especially when I've finally had a GPU that could run some of those ENB mods.
Posted on Reply
#99
BoxCutter
SithaerI'm curious to see how the final version will work cause its just too early to judge it in my eyes
Well nVidia themselves thought it's good enough to showcase and market, so it absolutely valid to judge it. People really need to understand that companies are not willing to publicly showcase something unless they think it looks good.
Posted on Reply
#100
Youseethis
Visible NoiseWut?

The devs are now wrong for creating the look they want?

Ok, you do you, I guess.
That's what you got from what I said?
They can do whatever they want. Use whatever tool they disire. I never said if they are right or wrong. And I was specific about the game: re9. They changed the conveyed meaning in that specific instance (Leon)

And that's because...
SithaerIn the Digital Foundry video I've watched they directly mentioned Todd Howard that he does in fact likes this DLSS 5 version of Starfield better than the original and he would prefer to play it like this
..Starfield is an upgrade. :roll:
SithaerPersonally I'm curious to see how the final version will work cause its just too early to judge it in my eyes, in some games it could be actually better imo if done right. 'Not that my 5070 will run this but thats another story :laugh: '
For know they said that they managed to run on a single 5090 instead of 2x 5090
Posted on Reply
Add your own comment
Nov 20th, 2026 15:45 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts