Nearly all graphics cards are expensive right now, but the ongoing RAM shortage is at least partly to blame for the current state of things. But, all shortages aside, few people would call Nvidia’s graphics cards cheap, and neither would I. They’re not, and they haven’t been for years (if ever).
However, are Nvidia’s GPUs truly overpriced? And even if they are, a better question is, are they bad value?
I believe that there’s one key aspect many people overlook when judging Nvidia’s GPU prices, and this has been the case for some time. We’re all forgetting about DLSS.
Graphics cards are just a vessel for Nvidia’s broader software stack
Let’s be real: Nvidia’s graphics cards were never exactly cheap, but the last few generations really took things up a few notches. And sure, we’ve had external factors to deal with that bled onto the pricing, such as the 2021-2022 GPU shortage, after which the prices were never the same again. But regardless of what caused it, the last three generations of Nvidia GPUs have come with sizeable price hikes, especially across the flagship models.
But Nvidia’s delivered some major improvements, too. And one of those major improvements was the addition of DLSS, which, in its current state, is a major selling point for Nvidia’s latest graphics cards. I’ve been calling Nvidia’s GPUs vessels for DLSS for a while now, and I stand by that opinion.
DLSS is groundbreaking because it changed what we even expect a graphics feature to do. It’s no longer just a clever way to squeeze out extra fps. Nvidia now treats DLSS as a full neural-rendering suite that can boost performance, improve image quality, and clean up some of the visual compromises that used to make upscaling feel like a last resort.
This is a huge deal, because it means your GPU no longer has to brute-force its way through every new game. The graphics card itself, as well as the game developers (for better or worse), now have a powerful crutch to lean on. DLSS gives players more room to keep a card comfortable for longer, whether that means using Super Resolution to make heavier games more manageable, Ray Reconstruction to improve ray-traced scenes, or newer DLSS 4.5 updates to get better image quality from the same basic starting point.
DLSS alone accounts for so many different aspects of owning a GPU, from future-proofing to actual performance, that it’s impossible to look at Nvidia graphics cards without accounting for how powerful DLSS has become.
The hardware behind DLSS has value outside of pure gaming
Those Tensor cores are more than just “fake frame” generators
Justin Duino / How-To Geek
That brings us to the hardware side of the equation, and specifically to Nvidia’s Tensor cores. They’re the reason DLSS exists in the first place, and also the reason why DLSS is limited to Nvidia GPUs. Reducing them to “fake frame” hardware is way too simplistic, though. Nvidia now treats them as the AI engine behind a much broader stack, with DLSS itself described as a suite of neural rendering tech instead of just a frame generation tool (which is what most gamers use it for).
That matters because the same hardware is now being put to work outside games, too. Nvidia is actively pushing RTX cards as local AI PCs, with support and optimization around things like on-device inference, RTX Video Super Resolution, Broadcast effects for voice and video, and other AI-assisted creative tools. Even if you buy the card for gaming first, you are still getting hardware that can accelerate a growing number of non-gaming workloads on the same machine.
The TL;DR here is that Nvidia’s specific hardware and the workloads it enables contribute to the high pricing. Even if your GPU is strictly a gaming tool, it’s still packed with hardware that makes certain things possible; things that wouldn’t work on an AMD or an Intel card due to the lack of the aforementioned hardware.
Let’s talk future-proofing
Especially important now, when hardware demands scale way too fast
DLSS, when used correctly, can turn a cheaper, older GPU into something much more equipped to face the future. And no, I’m not even talking about Nvidia CEO Jensen Huang’s claim that the RTX 5070 will beat the RTX 4090 at a third of the price.
The fact remains that without DLSS, we had two choices: either deal with our aging GPU by avoiding certain workloads or lowering in-game settings, or buy a new graphics card. But with DLSS, many GPUs are given a new lease on life. I don’t see any DLSS-powered GPU joining the list of obsolete Nvidia graphics cards anytime soon.
One caveat of DLSS is that it does need a reasonable base frame rate to be able to work its magic. So, you can’t magically turn a 10-year-old GPU into a blazing fast beast that can handle AAA games. But on cheaper and older graphics cards, it can make the difference between “unplayable” and “good enough.” Thanks to DLSS, many players can get away with skipping an extra generation or two when planning their upgrade path.
None of this changes the fact that the prices are high
And I’ve lost hope that they will ever get majorly better
DLSS is fantastic, and if I’m being honest, I believe that software-level upgrades are what we’re going to get more and more of. Hardware upgrades are slowing down, but DLSS evolves. Whether we like it or not, that price tag covers both the cost of the hardware and the cost of the software.
Yes, Nvidia graphics cards are expensive, or perhaps even overpriced, but when bought at MSRP, many models (not all) make sense to buy.
You have to give Nvidia some credit
At the risk of being called a shill, I do have to give Nvidia some credit here. It absolutely dominates the GPU market, and its GPUs are best in class if you want both the hardware and the software to align.
AMD can rival some of Nvidia’s cards in raster performance, but FSR 4 can’t rival DLSS 4.5 in the same way. Adoption is another key factor where Nvidia leads with ease.
With all that said, we shouldn’t have to pay between $650 and $800 for an RTX 5070. We shouldn’t even have to pay $550, but that’s a conversation that’s almost not worth having. This is never going to change for the better.
- Graphics RAM Size
-
12GB
- Boost Clock Speed
-
2600MHz
- Memory Bus
-
192-bit
Nvidia’s RTX 5070 graphics card is one of the more balanced options in the RTX 50-series lineup. It gives you 12GB of VRAM, which is much more future-proof than the 8GB VRAM lower-tier cards offer, while also offering more cores and higher bandwidth than the RTX 5060 Ti.
Stephan is the sports journalist for the Maple Grove Report.
