Sunday, April 5th 2026

Intel Texture Set Neural Compression Shrinks Textures by Up to 18x with Minimal Quality Loss

Intel has recently released a new video showcasing its latest Texture Set Neural Compression (TSNC) technology, which delivers textures up to 18 times smaller while maintaining visual quality with little to no noticeable difference compared to the industry-standard compression. Using AI-based neural networks, Intel's graphics team processes input data from industry-standard BCn textures. These textures are compressed through an AI model encoder, encoded in the latest space values, and then decoded by a network decoder to decompress the textures. The result is output data textures that are up to 18 times smaller, with some quality loss at maximum compression settings. As with any neural technology, TSNC is trained on millions of standardized textures to create an AI model that can replace traditionally compressed textures in the BCn format. This results in new, much smaller game textures that load faster, use less VRAM, and perform better thanks to modern GPU technology.

There are several ways to apply TSNC neural compression, depending on the desired outcome, whether it's saving game installation size, reducing VRAM usage, or improving performance. Variant A, as Intel calls it, can achieve up to 9 times compression of the standard texture set, with little to no difference in visual quality—almost an unnoticeable drop. However, when the goal is maximum efficiency and requires up to 18 times texture compression, Intel offers Variant B of the TSNC neural network. This variant provides a significant performance boost, with the trade-off being a modest visual change. Using NVIDIA's FLIP tool to measure quality drop in generated images, Intel notes that Variant A experiences a 5% visual quality drop, while Variant B sees up to a 7% quality drop, which is noticeably more.
You can judge for yourself by viewing the comparison images below.

Finally, Intel benchmarked the performance of its Texture Set Neural Compression technology using the latest "Panther Lake" system with Arc B390 integrated graphics, which also includes XMX cores to accelerate these technologies seamlessly. The AI model is producing the first texture pixel in about 0.194 nanoseconds, which is fast enough for users not to notice any additional latency or experience oddities in their game rendering. We can expect this technology to be shipped later this year as an alpha version, with beta and full stable releases expected later on, though no concrete timelines have been provided.
Sources: Intel Developers, via VideoCardz
Add your own comment

24 Comments on Intel Texture Set Neural Compression Shrinks Textures by Up to 18x with Minimal Quality Loss

#1
Hyderz
Is this similar to Nvidias neural texture compression?
Posted on Reply
#2
AleksandarK
News Editor
HyderzIs this similar to Nvidias neural texture compression?
Indeed it is, just Intel’s way.
Posted on Reply
#3
TumbleGeorge
I'm looking forward to games that take up as much storage as the recording of a few actual pixels and then "decompressed", or rather hallucinated by some LLM, to look like they have an actual terabyte of textures.
I'm thinking about another hyper-stupid competition focusing on just one of the software's features. From unnecessary big blowing to unnecessary great reduction.
Posted on Reply
#4
user556
It sounds a lot like a very large compression dictionary. It does beg the question of how big the graphics driver has become? And how big is this LLM model data?
Posted on Reply
#5
Chomiq
TumbleGeorgeI'm looking forward to games that take up as much storage as the recording of a few actual pixels and then "decompressed", or rather hallucinated by some LLM, to look like they have an actual terabyte of textures.
I'm thinking about another hyper-stupid competition focusing on just one of the software's features. From unnecessary big blowing to unnecessary great reduction.
I dunno about hyper-stupid competition when VRAM is getting more expensive, games use more of it and this tech is intended to reduce its usage. Variant A looks pretty good, with almost double the compression level of BCX and not that much of a variance from the source as variant B. I imagine that devs simply have to train the model on uncompressed textures, if that's the case the risk of hallucination should be decreased unlike with some general model that doesn't know what the source looks like.
Seeing how Nvidia announced their implementation of this couple of days ago, now Intel did it it's just a matter of time until AMD announces their implementation of the same concept.
Posted on Reply
#6
HLB
Corpos lie. They do that naturally, it's their modus operandi. So, when they say something like "minimal quality loss" it most likely means there is quality loss and they just assume few would notice or care. And they are probably right: very few would notice or care as a typical consumer is tasteless, ignorant and undemanding, they just want "more" and "cheap"... "quality" is low on most peoples' list.

So, buckle up as this will be getting worse now as every tech in the world will be getting the AI slop treatment. A handful of geeks and nerds will cry in horror while majority of the population will gulp the slop and ask for more with a smile.
Posted on Reply
#7
Tpanon
If there is quality loss, then I would argue it's not compression. It's reformatting into something else. Like turning a GIF into a JPEG. With real compression, you should be able to return to the original with nothing lost underway.
Posted on Reply
#8
RejZoR
I know for a fact that devs will misuse this just like they have tessellation, add 16Kx16K textures instead and for absolutely no reason cram so many of them on levels that even 16GB graphic cards will struggle while using neural textures. All with visual difference only visible when doing side by side pixel peeping comparisons.

Instead of doing the opposite, hyper optimizing games so they could run on basically anything. Isn't that the goal of making a game, so it can reach the greatest possible audience?

Tessellation had the same issue. Models already looked great, but they turned them into 5x polygons models for absolutely no reason whith most misc objects like railings and stuff remained stupid blocky. Where going the opposite direction and hyper optimizing games to remove even more polygons when not needed would make an amazing effect on performance.

So, yeah, sounds good on paper, I'm not holding my breath because devs surely will cock it up.
Posted on Reply
#9
cinemaware
HenryLootgraaba typical consumer is tasteless, ignorant and undemanding
HenryLootgraabA handful of geeks and nerds will cry in horror while majority of the population will gulp the slop and ask for more with a smile.
Insulting a large chunk of the population while suffering from delusions of grandeur is quite a mix, worthy of some French monarch ;)

The joke's on you though, and your fellow "geeks and nerds", since this lot gets it wrong all the time. This is a very good example: some AI tech that actually might be rather beneficial, with minimal cost, gets thrown into the "AI slop" bucket ...just because!
Posted on Reply
#10
HLB
cinemawareInsulting a large chunk of the population while suffering from delusions of grandeur is quite a mix, worthy of some French monarch ;)

The joke's on you though, and your fellow "geeks and nerds", since this lot gets it wrong all the time. This is a very good example: some AI tech that actually might be rather beneficial, with minimal cost, gets thrown into the "AI slop" bucket ...just because!
Ouch, ouch, ouch... hit a nerve? I am often the the first one to defend average users accused of "skill issues" but there is no denial that majority of the world population is tasteless and clueless. It has been for thousands of years. It's just the way the world ticks.
Posted on Reply
#11
zapster
This & similar tech problems is easily solved if the devs of games can control exactly the output/what you see, the final outcome, then there is not much transformed.

If the devs program & design the final outcome & decide exactly it would be less heresy than say when bluray players upscaled DvD which I have never heard of any whining about.

Upscale yes transform decided by dev 100%.
Posted on Reply
#12
TheinsanegamerN
RejZoRInstead of doing the opposite, hyper optimizing games so they could run on basically anything. Isn't that the goal of making a game, so it can reach the greatest possible audience?
Based on your logic games should have never left the text based stage, since literally anything can run those, right? There's nothing more optimized then TEXT. Why would you use 2d sprites, you're not optimizing for the commodore VIC-20 crowd!

No matter how optimized a product, people will still cry because it doesnt run on their ATI HD 4250 or whatever old silicon they found in a dumpster. At some point, you need to upgrade your hardware for modern requirements. Optimization is balanced with features and visual quality, always has been.
HenryLootgraabOuch, ouch, ouch... hit a nerve? I am often the the first one to defend average users accused of "skill issues" but there is no denial that majority of the world population is tasteless and clueless. It has been for thousands of years. It's just the way the world ticks.
People said the same of DLSS upscaling when it came out, now look at the magic DLSS4+Transformer is capable of.
Posted on Reply
#14
RejZoR
TheinsanegamerNBased on your logic games should have never left the text based stage, since literally anything can run those, right? There's nothing more optimized then TEXT. Why would you use 2d sprites, you're not optimizing for the commodore VIC-20 crowd!

No matter how optimized a product, people will still cry because it doesnt run on their ATI HD 4250 or whatever old silicon they found in a dumpster. At some point, you need to upgrade your hardware for modern requirements. Optimization is balanced with features and visual quality, always has been.

People said the same of DLSS upscaling when it came out, now look at the magic DLSS4+Transformer is capable of.
When games run like dog shit on RX 9070 XT then we absolutely need to have a chat. And they don't look anything spectacular or jaw dropping. That's why. I never said they shouldn't evolve, what I did say was evolving meaningfully. If engine can look the same as all top games but have absurdly low requirements, isn't this evolution in itself? Game studios don't give a shit if their game has most market share on NVIDIA's flagship graphic cards only. They care they can sell the most games. Games that look great and run great on anything have always been highly praised.
Posted on Reply
#15
TheinsanegamerN
RejZoRWhen games run like dog shit on RX 9070 XT then we absolutely need to have a chat. And they don't look anything spectacular or jaw dropping. That's why.
Crysis ran poorly on hardware of its day. That was 20 years ago. There will always be games that run badly.
RejZoRI never said they shouldn't evolve, what I did say was evolving meaningfully.
Neither "evolve" or "meaningfully" were anywhere in your comment.
RejZoRIf engine can look the same as all top games but have absurdly low requirements, isn't this evolution in itself? Game studios don't give a shit if their game has most market share on NVIDIA's flagship graphic cards only. They care they can sell the most games. Games that look great and run great on anything have always been highly praised.
You assume that you can get the same level of graphical fidelity on "absurdly low requirements. That in itself is vague, what is absurdly low? What you consider absurd others would describe as obsolete, or what they consider absurd you describe as powerful. How many features should we hold back to promote old hardware?

RE: Requiem is well optimized, but RT performance still requires modern hardware to run well even at 1080p. Should RT and PT be removed so people still using RTX 2060s feel better about being able to run the game? What about Crimson Desert? Both sold very well despite not being designed to "sell the most games" as you say.

Where does the buck stop? How much graphical fidelity should be sacrificed to achieve better performance? Thats what I was getting at. What makes your baseline the acceptable baseline?
Posted on Reply
#16
swaaye
This stuff is interesting but I get the impression that current GPUs are rapidly losing value. Between this neural compression and the upscaling techs. Don't spend much on a card because all of them are going to have a very short shelf life.
Posted on Reply
#17
RejZoR
TheinsanegamerNCrysis ran poorly on hardware of its day. That was 20 years ago. There will always be games that run badly.

Neither "evolve" or "meaningfully" were anywhere in your comment.

You assume that you can get the same level of graphical fidelity on "absurdly low requirements. That in itself is vague, what is absurdly low? What you consider absurd others would describe as obsolete, or what they consider absurd you describe as powerful. How many features should we hold back to promote old hardware?

RE: Requiem is well optimized, but RT performance still requires modern hardware to run well even at 1080p. Should RT and PT be removed so people still using RTX 2060s feel better about being able to run the game? What about Crimson Desert? Both sold very well despite not being designed to "sell the most games" as you say.

Where does the buck stop? How much graphical fidelity should be sacrificed to achieve better performance? Thats what I was getting at. What makes your baseline the acceptable baseline?
That's a false equivalency. Crysis didn't ran poorly because it was poorly made or anything. It was so far ahead of the industry that it made it run poorly. And to be honest, the whole "can it run Crysis" has been dramatically overblown. Game scaled really well depending on settings and even though they made the "remaster", the remaster actually looks dumber and more basic than the original. It's only in few places where new one looked better and it was mostly just increased view range beyond what game offered in settings. A mod would probably allow that. The actual effects, visuals and physics were just better in original.

Hell, original Far Cry is ancient at this point and it still looks pretty despite being dated now. Same with Crysis.
Posted on Reply
#18
zus
Using NVIDIA's FLIP tool to measure quality drop in generated images, Intel notes that Variant A experiences a 5% visual quality drop, while Variant B sees up to a 7% quality drop, which is noticeably more.
I suppose they're comparing them against uncompressed textures, right? I wonder how BCn performs then, is TSNC actually beating it in quality?
TpanonIf there is quality loss, then I would argue it's not compression. It's reformatting into something else. Like turning a GIF into a JPEG. With real compression, you should be able to return to the original with nothing lost underway.
Lossy compression isn't new. In fact, JPEG is a lossy compression algorithm (unlike PNG, which is lossless).
en.wikipedia.org/wiki/Lossy_compression
Posted on Reply
#19
evernessince
ChomiqI dunno about hyper-stupid competition when VRAM is getting more expensive, games use more of it and this tech is intended to reduce its usage. Variant A looks pretty good, with almost double the compression level of BCX and not that much of a variance from the source as variant B. I imagine that devs simply have to train the model on uncompressed textures, if that's the case the risk of hallucination should be decreased unlike with some general model that doesn't know what the source looks like.
Seeing how Nvidia announced their implementation of this couple of days ago, now Intel did it it's just a matter of time until AMD announces their implementation of the same concept.
AMD already did back in 2024: gpuopen.com/download/2024_NeuralTextureBCCompression.pdf
TpanonIf there is quality loss, then I would argue it's not compression. It's reformatting into something else. Like turning a GIF into a JPEG. With real compression, you should be able to return to the original with nothing lost underway.
Lossy compression is a thing. I think it's fine so long as it's not changing the texture every time you look away and then back and if the quality can match other high quality lossy compression methods.
TheinsanegamerNPeople said the same of DLSS upscaling when it came out, now look at the magic DLSS4+Transformer is capable of.
Yes and the criticism was warranted and needed in order to get where we are today. If people just accept garbage, there's no incentive for companies to improve. You would not have the DLSS you have today if it wasn't for the critics.
RejZoRThat's a false equivalency. Crysis didn't ran poorly because it was poorly made or anything. It was so far ahead of the industry that it made it run poorly. And to be honest, the whole "can it run Crysis" has been dramatically overblown. Game scaled really well depending on settings and even though they made the "remaster", the remaster actually looks dumber and more basic than the original. It's only in few places where new one looked better and it was mostly just increased view range beyond what game offered in settings. A mod would probably allow that. The actual effects, visuals and physics were just better in original.

Hell, original Far Cry is ancient at this point and it still looks pretty despite being dated now. Same with Crysis.
Yeah Crysis 1 was really clean looking and the environment is more destructible / intractable than many AAA games today. Very few games nowadays aren't filled with noise or temporal artifacts chasing realism but they forget they art direction is more important than realism.
Posted on Reply
#20
_roman_
i think that answers partly my thought

companies rediscovered compression

what a surprise. i also rediscovered compression for my file systems and years ago for the kernel.

factor 20 is the usual. which data is dropped? which details are dropped? How much extra instruction cycles do you need for that "neural fancy word bla bla ai fancy texture compression method company a b c d" ?

i agree, tehy may have discovered lossy compression.

wtf

we discuss now bitmaps?

msi uefi generates bmp when i create screenshots as the settings can not be exported!

1.5 mb per file which are 0.05mb when converted for 10xx x xxx resolution most likely 1024x768
note i talk about the (BMP) on the left.

LOL intel discovered bitmaps. haha windows 95 bitmaps.



example, not the same file contents, but 2 original msi uefi files vs converted quick with gimp most likely files
note: i did not care which dataformat jpeg png, whatever, just smaller at that time.

2.4 MB / 0.3 MB = factor 8

Posted on Reply
#22
chrcoluk
swaayeThis stuff is interesting but I get the impression that current GPUs are rapidly losing value. Between this neural compression and the upscaling techs. Don't spend much on a card because all of them are going to have a very short shelf life.
I think it has the potential to accelerate even further low VRAM GPU's assuming they dont get the tech.
High VRAM GPUs might be ok as they will have the hardware to cope with uncompressed still.
Posted on Reply
#23
cvaldes
evernessinceYes and the criticism was warranted and needed in order to get where we are today. If people just accept garbage, there's no incentive for companies to improve. You would not have the DLSS you have today if it wasn't for the critics.
This is not specific to DLSS or computer graphics technology. This applies to pretty much EVERYTHING people do.

If people didn't care, we would probably just be riding horses. There would be no Teslas, Lucids, BYDs on the roads. And yeah, that lunch you just finished would look a lot like pig swill or something your cat might not eat. And yeah, you wouldn't be reading this site on a big flat panel display connected to a powerful desktop computer (or even viewing it on your smartphone).

Ultimately consumers set the expectations of what they find desirable. And not everyone is the same everywhere.

Here in the USA, public transit is basically a joke. In Japan and Switzerland, public transit is run by mature, responsible and accountable adults. I'm most familiar with the various transit agencies here in the SF Bay Area and sadly our agencies are NOT operated the same way. Because the vast majority of consumers don't give a sh!t.

DLSS had an especially rough time at its debut because AMD and Intel didn't have anything to challenge Nvidia.

Hell, McDonald's in Japan blow doors on their American counterparts. Because the consumers care.

If no one complained about DLSS 1.0, that's all there would be in 2026.

And your smartphone battery would be dead in 4 hours. And it wouldn't survive a dunking. And the camera would be shooting 640x480p images.
Posted on Reply
#24
Vayra86
TpanonIf there is quality loss, then I would argue it's not compression. It's reformatting into something else. Like turning a GIF into a JPEG. With real compression, you should be able to return to the original with nothing lost underway.
Wowowow you are asking for actual precision and controllability in a computer environment? Mate... we didn't burn 80 billion for AI to be precise! Such expectations... Ballpark accuracy is much more profitable! We'll deliver it at 4K so we can sell you the 'fact' it should be high fidelity, ok?
Posted on Reply
Add your own comment
Nov 21st, 2026 04:29 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts