If you haven't bought a new GPU in the last 5 years, you will find the market a bit different from what you remember. While you might expect high-end GPUs to feature high-end specs in every department, the reality is that manufacturers don't always agree. Whether you call it sandbagging or willful ignorance, VRAM on GPUs has been hit the hardest in recent years. Even $1,000 GPUs aren't immune to the plague of insufficient VRAM in 2025.

The concerns around low VRAM that started brewing around 2020 have finally come to a head. Nvidia is still acting stingy about VRAM on every SKU except the flagship. More and more games are starting to saturate 8GB framebuffers at 1080p, and high-resolution gaming is in dangerous waters, thanks to just 12GB VRAM even on $500+ GPUs. AMD and Intel are faring considerably well when it comes to VRAM capacity, but we need more accountability from GPU manufacturers overall.

5 Expect stutters and crashes with insufficient VRAM

Generous VRAM is a non-negotiable

The whole furor surrounding GPU VRAM isn't for nothing. Your GPU needs enough VRAM to unpack game textures, access the framebuffer (often used synonymously with VRAM), and process other graphical information to render games efficiently. As texture sizes increase, and high-res gaming becomes more popular, games (and creative applications) utilize increasingly high amounts of VRAM. Running out of VRAM is never going to end well.

This is because the moment your GPU runs out of VRAM, the game you're running will experience incomplete textures, stutters, or even crashes. Without sufficient VRAM, the other components of your PC will not be able to pick up the slack, no matter how high-end they are. In the past, VRAM was rarely an issue, since most games weren't demanding enough to saturate the framebuffer of most GPUs. In 2025, however, we have enough titles capable of doing just that.

4 Games are getting VRAM-hungry, even at 1080p

Budget gamers might have to avoid the latest games

Over 50% of gamers still use 1080p as their primary resolution, followed by 1440p at around 30%, and 4K at around 3%. While the most popular resolution used to be immune from the woes of insufficient VRAM around five years ago, things have changed drastically now. Suppose you're running a particularly intensive game with an RTX 3070 on your 1080p monitor. In most instances, the 8GB VRAM will be enough to tackle the game's requirements.

However, some titles, such as Cyberpunk 2077, Alan Wake 2, The Last of Us Part 1, and Avatar: Frontiers of Pandora, can easily overcome the RTX 3070's 8GB framebuffer at high settings, even without enabling ray tracing. 8GB on the RTX 3070 was highly questionable even at launch, considering the GPU was quite powerful on almost every other parameter. Nearly 5 years later, GPUs like the RTX 3070 are facing the brunt of the VRAM fiasco, and ray tracing is piling on the requirements.

With the way games have been progressing since 2023, 8GB of VRAM is only justifiable on budget GPUs. Mid-range GPUs should feature a minimum of 16GB VRAM, with 20GB+ of VRAM being the ideal on high-end models. 1440p and 4K gaming are already pushing 16GB of VRAM usage in some of the most demanding titles (with RT and frame generation enabled), hence gamers buying a new GPU today are forced to opt for expensive models that have enough VRAM for their needs.

3 Many titles are implementing full ray tracing

​​​VRAM requirements are getting crazier by the day

Years ago, full ray tracing used to be unimaginable even on modern GPUs. In 2023, it made its way to Cyberpunk 2077 as an update, and was present in Alan Wake 2 from the ground up. Full ray tracing is a more comprehensive form of ray tracing that traces every light source in the game to enable ray-traced global illumination, and more realistic shadows and reflections. All this wizardry comes at a computational cost that demands more VRAM than ever.

At the moment, the number of titles utilizing this technique is low, but it won't stay that way. Assassin's Creed: Shadows doesn't allow you to disable ray tracing completely, and upcoming games like Half-Life 2 RTX and Doom: The Dark Ages are set to feature full ray tracing as well. Older GPUs with poor ray tracing performance in addition to low VRAM are being left behind by the industry, and gamers need to demand larger VRAM capacities on the latest models.

2 High-end GPUs with low VRAM will suffer the most

16GB on $1,200 GPUs is a joke

To some extent, entry-level and budget GPUs having less-than-ideal VRAM is excusable, since cramming them with VRAM isn't going to make up for their poor raw performance. High-end GPUs, on the other hand, need to be supported with a framebuffer large enough to not become an unfortunate bottleneck. If I'm investing $800–$1000 in a GPU, I don't want to worry about running out of VRAM.

Sadly, that's exactly the case with many high-end GPUs these days. Older models like the RTX 4070 Ti (12GB) and RTX 4080 Super (16GB), and recent ones, such as the RTX 5080 (16GB), feature terribly low VRAM for their price points. Even if these powerful GPUs are capable of driving high FPS at high settings in the most demanding games, their VRAM capacities will enforce an artificial ceiling on performance. And, unfortunately, you can't add more VRAM to your GPU.

1 Manufacturers need to do better in 2025

All eyes are on Nvidia

AMD and Intel have kept pace with the demand for more VRAM over the last few years, but Nvidia has refused to budge. When AMD's RX 6000 and RX 7000 series offered 16GB and 24GB of VRAM on models cheaper than the competition's, and Intel's Arc Battlemage SKUs did the same, how can Nvidia continue to skimp on VRAM? Team Green enjoys a 90% market share of the GPU market, and gets away with almost anything, whether it's poor VRAM, burning connectors, black screens, or missing ROPs.

It's bad enough that Nvidia's RTX 50 series offers disappointing gains over the RTX 40 series, but combine that with the company's stance on VRAM, and you get a concoction that no one wants. It's not even expensive to add more VRAM to a GPU; Nvidia's sole reason for offering enough VRAM only on its flagship model can be seen as sandbagging — forcing those who care enough about VRAM to opt for the most expensive SKU.

Consumers need to hold companies accountable, making a bigger hue and cry over insufficient VRAM than we see right now. It might not do much at first, but Nvidia just might do better in the next generation. After all, the RTX 5070 Ti featured 16GB of VRAM, a significant increase compared to the 12GB seen on its predecessor.

The time to demand more VRAM is now

It's disheartening that the community needs to beg manufacturers for something that should be present on modern GPUs by default. The reality, however, deems it necessary that we, as consumers, start making our feelings known, and reject GPUs with poor VRAM capacities. Meanwhile, if you're in the market for a new GPU, steer clear of models with less than 12GB of VRAM. And if you're gaming at 1440p and 4K, try to target GPUs with 16GB of VRAM and more.