Thursday, August 6th 2026

NVIDIA Neural Texture Compression Now Runs on RTX Spark

NVIDIA has officially brought its RTX Neural Texture Compression (NTC) technology to Windows-on-Arm, gearing up for a general release on its RTX Spark PC platform. First demonstrated at GTC 2026 in May, NVIDIA showcased how NTC can significantly reduce GPU VRAM usage by up to seven times. In a technology demo at one of the GTC 2026 sessions, NVIDIA revealed that NTC can lower VRAM usage from 6.5 GB to just 970 MB in certain scenes. NTC technology is an AI-driven texture output that enables games to feature high-quality, complex materials without sacrificing performance. Games can take advantage of the substantial VRAM reduction that NTC offers while maintaining the same texture quality.

For the upcoming RTX Spark, NVIDIA is enhancing the software ecosystem so gamers and future buyers receive a mature software that works seamlessly out of the box, eliminating the need for workarounds. Even Japan's legendary SEGA has released AAA titles optimized for RTX Spark and is dedicating more resources to support the ecosystem. NVIDIA has already started shipping its first preview build of the CUDA toolkit for the RTX Spark platform, specifically as a native Windows-on-Arm software toolkit. Now, emerging technologies like NTC will run on RTX Spark, which includes configurations such as one with 6,144 CUDA cores and another with 5,120 CUDA cores. Both configurations are based on the "Blackwell" GPU microarchitecture and are paired with a 20-core Arm CPU and up to 128 GB of LPDDR5X memory.
Traditionally, games use block-compressed formats like BCn, such as BC5, BC6, or BC7, which are commonly applied in 4x4 pixel formats and rendered by the GPU. However, NVIDIA has trained small neural networks that can produce the desired pixel format and texture appearance at a fraction of the size of traditional texture compression formats. Instead of using gigabytes of VRAM for textures, NTC drastically reduces VRAM usage by emulating textures, allowing for either much lower VRAM consumption or significantly enhanced material appearance.
Sources: NVIDIA on GitHub, via VideoCardz
Add your own comment

17 Comments on NVIDIA Neural Texture Compression Now Runs on RTX Spark

#1
Jyubei
Very promising tech, but we'll see how long it takes to get to the gaming space.

If it's only on the DX Spark now, then this may take a while to get to us.

Ideally it would be an open source API tech, embedded in DX12, so that us AMD folks could use it as well... Alas, gotta wait and wait longer.
Posted on Reply
#2
Blazeoptimus
JyubeiVery promising tech, but we'll see how long it takes to get to the gaming space.

If it's only on the DX Spark now, then this may take a while to get to us.

Ideally it would be an open source API tech, embedded in DX12, so that us AMD folks could use it as well... Alas, gotta wait and wait longer.
Nvidia is typically very guarded with its IP. I expect that if we ever see this in an AMD or Intel product it will be a competing standard that's close, but not quite the same quality level of Nvidia's proprietary standard - and that's the point. Nvidia has largely cornered the market with such tactics.
Posted on Reply
#3
mikesg
Why is SEGA working on games supporting this.. are they going to make a new Game Gear based on the RTX Spark that costs $5,000?
Posted on Reply
#4
Vya Domus
Blazeoptimuscompeting standard that's close, but not quite the same quality level of Nvidia's proprietary standard
Proprietary =/= best.
Posted on Reply
#5
son_of_wasps
Does neural texture compression require developer support, or is it application agnostic? I.e., do the developers need to pre-compress the textures, or can they be compressed on-the-fly by the CPU or GPU before storage in VRAM?

If compression can be done on-the-fly, this would enable Nvidia to dramatically reduce vram on their cards (e.g. selling 4gb cards again) thus reducing cost and undercutting competitors all while providing the same level of visual fidelity.

Bonus points if compression can be precompiled then cached on disk to avoid stuttering when new scenes are loaded.
Posted on Reply
#6
Bigshrimp
JyubeiVery promising tech, but we'll see how long it takes to get to the gaming space.

If it's only on the DX Spark now, then this may take a while to get to us.

Ideally it would be an open source API tech, embedded in DX12, so that us AMD folks could use it as well... Alas, gotta wait and wait longer.
They will lock it into their product line. They won't let it be open source, just like all of the rest of the products they spend R&D on. That means it will probably only ever work on Nvidia's products until AMD and others make their own.
Posted on Reply
#7
zus
I wish they could put this amount of effort on their Linux driver too, oh well.
son_of_waspsIf compression can be done on-the-fly, this would enable Nvidia to dramatically reduce vram on their cards (e.g. selling 4gb cards again) thus reducing cost and undercutting competitors all while providing the same level of visual fidelity.
Intel is already working on their own solution. AMD will probably have their own version of the tech too, so I don't think NVIDIA will have this kind of advantage, at least not for too long.
Posted on Reply
#8
Blazeoptimus
BigshrimpThey will lock it into their product line. They won't let it be open source, just like all of the rest of the products they spend R&D on. That means it will probably only ever work on Nvidia's products until AMD and others make their own.
If you think about it, from a monetary standpoint, Nvidia's approach as been far and away the better approach. Take AMD's FSR vs Nvidia DLSS. Until the latest versions, FSR ran on anything, including Nvidia cards. This lead it to be widely used in games. However, DLSS was always reviewed as superior to FSR. What this lead to was a consumer preference for Nvidia hardware, since even if the games they played didn't support DLSS, FSR would still run on the nvidia hardware without issue. The opposite was not true. The only place where 'anti open source' mentality has cost Nvidia is in the linux desktop market. In that market, AMD is generally the preffered option. However, linux desktop accounts for such a small portion of the total market (even if you include steam devices), its has had no meaninful impact on either Nvidia or AMD.
Posted on Reply
#10
evernessince
BlazeoptimusIf you think about it, from a monetary standpoint, Nvidia's approach as been far and away the better approach. Take AMD's FSR vs Nvidia DLSS. Until the latest versions, FSR ran on anything, including Nvidia cards. This lead it to be widely used in games. However, DLSS was always reviewed as superior to FSR. What this lead to was a consumer preference for Nvidia hardware, since even if the games they played didn't support DLSS, FSR would still run on the nvidia hardware without issue. The opposite was not true. The only place where 'anti open source' mentality has cost Nvidia is in the linux desktop market. In that market, AMD is generally the preffered option. However, linux desktop accounts for such a small portion of the total market (even if you include steam devices), its has had no meaninful impact on either Nvidia or AMD.
What you are describing is a strategy that systemically advantages the largest player in a market while disadvantaging everyone else. In the case of using proprietary technologies, Nvidia has the capital and manpower advantage. It also has a decade of entrenchment during the 8 years it was building CUDA while AMD was going bankrupt. When Nvidia pushes these options, it hurts choice for games. It doesn't just mean you can't use them on AMD graphics cards, it also discourages new players in the market. Instead of just making the best GPU possible, new GPU players have to worry about an increasing pile of games with features that only work on a single vendor's cards. So now they have to deal with a huge value imbalance where your cards simply cannot have features in all these games unless you work with the devs on updating them (assuming the devs are even still around). The longer we have these proprietary technologies and the more abudent, the greater the technical debt any new market player will be in and the more work competitors need to do to keep up. It instantly puts your competitors on the backfoot as the leading incumbent. If everyone is always busy playing catchup with the leader, the leader always wins.

You almost word it in a way that implies AMD could have gone proprietary only a decade back and saw increased success when in fact open source was probably the only option that was financial viable for them and that didn't eat up what little engineering resources they had. Realistically AMD didn't have the choice of going proprietary until after it saw success with a few generations of Zen and had money coming in. Even then, AMD going prorietary isn't a good thing for market. It's AMD trying to play Nvidia's game, they are doing so at a significant resource and implementation disadvantage as disucessed earlier and it's bad for customers. Either way you slice it, Nvidia is dictating the market.

DLSS wasn't what started Nvidia being percieved as higher value. That started far far before DLSS ever came around. Even Fermi outsold it's AMD counterparts 3 to 1 and that was ages before DLSS was even a twinkle in Jensen's eye.

NTC being proprietary is certainly not a good thing either but one can hope that any impelementations will go through an API middle layer that will allow game devs to route that to vendor appropriate technologies. That said, if path tracing implementations are anything to go by and their lack of support in some games for AMD hardware despite the API and hardware supporting it, I cannot say I'm hopeful.
Posted on Reply
#11
DavidC1
What's promising about this tech? The AI slop screenshot with games they had was part of the whole Neural Rendering thing.

Aside from further perpetuating the surveillance datacenter grid, making every electronics more expensive, this will also ruin games.
Posted on Reply
#12
JustBenching
evernessinceWhat you are describing is a strategy that systemically advantages the largest player in a market while disadvantaging everyone else. In the case of using proprietary technologies, Nvidia has the capital and manpower advantage. It also has a decade of entrenchment during the 8 years it was building CUDA while AMD was going bankrupt. When Nvidia pushes these options, it hurts choice for games. It doesn't just mean you can't use them on AMD graphics cards, it also discourages new players in the market. Instead of just making the best GPU possible, new GPU players have to worry about an increasing pile of games with features that only work on a single vendor's cards. So now they have to deal with a huge value imbalance where your cards simply cannot have features in all these games unless you work with the devs on updating them (assuming the devs are even still around). The longer we have these proprietary technologies and the more abudent, the greater the technical debt any new market player will be in and the more work competitors need to do to keep up. It instantly puts your competitors on the backfoot as the leading incumbent. If everyone is always busy playing catchup with the leader, the leader always wins.

You almost word it in a way that implies AMD could have gone proprietary only a decade back and saw increased success when in fact open source was probably the only option that was financial viable for them and that didn't eat up what little engineering resources they had. Realistically AMD didn't have the choice of going proprietary until after it saw success with a few generations of Zen and had money coming in. Even then, AMD going prorietary isn't a good thing for market. It's AMD trying to play Nvidia's game, they are doing so at a significant resource and implementation disadvantage as disucessed earlier and it's bad for customers. Either way you slice it, Nvidia is dictating the market.

DLSS wasn't what started Nvidia being percieved as higher value. That started far far before DLSS ever came around. Even Fermi outsold it's AMD counterparts 3 to 1 and that was ages before DLSS was even a twinkle in Jensen's eye.

NTC being proprietary is certainly not a good thing either but one can hope that any impelementations will go through an API middle layer that will allow game devs to route that to vendor appropriate technologies. That said, if path tracing implementations are anything to go by and their lack of support in some games for AMD hardware despite the API and hardware supporting it, I cannot say I'm hopeful.
Great in theory, irrelevant in practice. Most nvidia sponsored games had FSR. Most amd sponsored games didn't have DLSS. Also since most games also run on consoles that have AMD hardware AND use upscaling, it's nvidia that's at a disadvantage.

Open implementations mean that someone is going to spend millions in R&D for something their competitors can use.
Posted on Reply
#13
Grimling
notoperableCan it compress the price tag?
Of course it can!
Get ready for the RTX 6090 8GB! for only 2999 USD (MSRP, real market price may vary).
That is a Price reduction by more than 100% compared to the real market price of the 5090 today!
100% cheaper, price compression complete.
Posted on Reply
#14
Dukenukemx
Cool idea, except AMD and Intel are also doing it. We don't need another upscaler situation where every hardware manufacturer has their own solution. I'd also like to see this tech be used to reduce file sizes because games have gotten far too big lately, but if AMD, Intel, and Nvidia have their own version of neural texture compression then that's not going to be possible.
Posted on Reply
#15
Razor512
Their new texture compression would heavily depend on how well they are able to extend support to older Gen cards. If it becomes a feature that needs the latest GPU. For example, if it becomes a feature that only 1% of gamers can use, then a game studio will be unlikely to dedicate resources towards implementing it, since they will not want to alienate or compromise the experience for the other 99%.
Posted on Reply
#16
evernessince
Razor512Their new texture compression would heavily depend on how well they are able to extend support to older Gen cards. If it becomes a feature that needs the latest GPU. For example, if it becomes a feature that only 1% of gamers can use, then a game studio will be unlikely to dedicate resources towards implementing it, since they will not want to alienate or compromise the experience for the other 99%.
Nvidia will probably be using FP4 or BF4 to ensure it runs performant on 5000 series with an 8 bit fallback. So 5000 and 4000 series and newer only.

NTC will also require devs to compress an additional set of their textures in a specialized format, so it'll add to a game's size. Devs would basically be compressing textures just for newer Nvidia cards without a standard API implementation that works with other vendors. Nevermind that I'm not certain it'd run on lower end cards that must also be running DLSS and other tech that uses the tensor cores like the AI denoiser for RT / PT.

I can imagine some games might implement it, simply because Nvidia can sponser them to do so. They have a ton of money to throw around.
Posted on Reply
#17
bug
Both configurations are based on the "Blackwell" GPU microarchitecture and are paired with a 20-core Arm CPU and up to 128 GB of LPDDR5X memory.
The obvious problem being, the only customers for these will be likes of Mr. Zuckerberg, Musk or Huang himself.
I've been building my own PCs foe like 3 decades and I don't recall the PC ever being in the bad position we see today 😭
Posted on Reply
Add your own comment
Nov 22nd, 2026 06:05 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts