Club Benefits
For the first time in a long while, Nvidia GTC 2026 gave us some huge PC gaming news. CEO Jensen Huang introduced DLSS 5: the “fusion of 3D graphics and artificial intelligence” in his own words, and the response has been…mixed to say the least.
From outlets calling it the most impressive tech they’ve seen in a long time to people throwing out AI slop critiques and comparing it to those face filters you get with certain smartphone cameras, opinions are all over the place. And honestly, I didn’t know how to feel at first — like that damn dress meme in 2015, my mind changed every time I looked at the comparison videos!
So I did what I always do. I followed my late Grandad’s advice to “sleep on it, because you’ll know how you feel in the morning.” And in short, DLSS 5 is absolutely a breakthrough, but game developers have until the fall to find the sweet spot for it in their games. Let me explain.
Article continues belowWhat is DLSS 5? Let’s explain it with pizza
It feels literally like yesterday that I was talking about DLSS 4.5, but we’re already marching onto the next iteration — and it’s a big one we’ve known was coming for a while now.
“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,” Huang commented.
It’s a fusion of the “predictive” model that’s fueled DLSS for a while now with the “probabilistic” elements of generative AI to bring photorealism to games with cinematic lighting, enhanced material depth, real-time neural rendering and temporal consistency. And all of its capabilities are controllable by the game developers, as they can tune the intensity, color and masking to find an enhancement balance that is right for them.
To break this down, I’m feeling hungry and you know what that means… Back to the pizzeria I go!
Think about previous versions of DLSS as a magic magnifying glass. I love that 8-inch pizza, but I want more of it, so DLSS takes that pizza and stretches it to look like a 16-inch XL — making it bigger without the crust getting too thin or the cheese burning more easily (this is resolution scaling).
Then every now and again, the chef slides an extra few slices into the box between every real slice you actually order, so it feels like you’re eating more at a faster pace (frame generation).
Now, with the fifth generation, there’s an AI master chef that doesn’t just stretch the pizza, it re-imagines it. With settings turned up to max, the chef looks at that cheap pepperoni slice and says “I know what you were trying to do here, but I can do better,” and swaps it with artisanal, hand-cured salami and fresh buffalo mozzarella — even though you didn’t order those things.
Basically, DLSS 5 has stopped trying to “show you the game better,” and is now resorting to “showing you a better version of the game.”
Bridging the uncanny valley
We can’t say any of us didn’t see this coming. Jensen himself talked about it at CES 2026 in a behind-closed-doors Q&A session and shot his shot at a growing “fusion between rendering and generative AI.”
“In the future, it is very likely that we'll do more and more computation on fewer and fewer pixels. By doing so, the pixels that we compute are insanely beautiful, and then we use AI to infer what must be around it.” Huang said. He talked about the “utterly shocking and incredible” results he saw in the labs that looked like “basically a photograph interacting with you at 500 frames per second.”
And now we have our first glimpse at it. In some of the game videos it’s a significant improvement, but in others you can start to spot some creative challenges that will surely be worked out as we close in on an official launch.
Let's start with the faces (yep, we’ve got to talk about Grace’s face in Resident Evil: Requiem). There was a recent brain scan study that showed our brains process “hyper-realistic AI faces” differently than real faces. A sudden spike in activity around 600ms after seeing an image that triggers an internal uneasy mismatch feeling.
That’s what I believe is happening here, and a lot more of it comes down to the cinematic lighting than I initially thought. Based on my time pixel peeping videos and image comparisons, I’d say about 60% of Grace’s face tweaks here can be explained by lighting and material depth.
The remaining 40% is neural rendering — there are definitely fuller lips and sharper jawlines (not that Leon Kennedy needs it, being the smokeshow he is).
In some places, this photorealism really shines — games like EA Sports FC and Starfield truly benefit from this upgrade. But I can appreciate the view around creative interpretation vs AI creating a jarring effect.
There are some other things I noticed, too. Surfaces and textures have been given a serious upgrade, but with DLSS’ reinterpretation of lighting, some of it feels less stylized or moody.
Take this scene from Nvidia's Zorah Demo, for example. Before, there was a warmer hue and intentional shadowing to add depth, but that is re-interpreted with the cinematic lighting in a way that I feel loses the vibe a little.
All-in-all, DLSS 5 is a diamond in the rough. I can see what the intention is, but this is all completely in the hands of devs to use however they wish. Maybe Capcom rolls back on the face tech, or Warner Bros. alters the cinematic lighting. It will take time to find the right balance.
I’ve got some questions
But of course, this is one PC gamer who’s tested all the best GPUs and the tech that enables them — pixel peeping videos and screenshots. I think it’s something we all have to interact with to get a fuller understanding of DLSS 5.
And if Nvidia’s reading this (hi btw), before I (hopefully) get some hands (and eyes) on time with it, I do have some questions to get a better understanding of what’s going on under the hood:
- What is that real-time neural rendering and what has it been trained on?
- How big is the DLSS 5 model now, and what is the goal for it by fall? Currently demos showed on 2 RTX 5090s — so optimization/compression is key.
- What are the controls for more stylized games? Titles that don't necessarily benefit from photorealism with cel-shaded or more artistic graphics.
DLSS5 outlook
The timing of these announcements always makes me chuckle. I know it may be unintentional, but it just feels like once another GPU company announces an update to its upscaling/frame generation tech, Nvidia just says “hold my beer” and takes another giant step ahead of the pack.
Intel XeSS 3 with multi-frame generation? How about a little DLSS 4.5. AMD FSR Diamond? Well, here’s DLSS 5. And honestly, it is a breakthrough — but I can totally understand the mixed response.
In some titles, it is a generational leap forward. In others, it can feel like an AI veneer. But ultimately, the tech is here and devs can turn it up or down however they want. Nvidia isn’t adding slop to games, that team is marching forward to photorealism with another feather in the cap of developers to bring their visions to life.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
