Graphics Playing cards Might Get Even Extra Costly, and It’s Not Simply Nvidia

Have you already given up on trendy GPU costs? If not, you could quickly don’t have any alternative however to just accept that this era of graphics playing cards is totally unaffordable. As if PC gaming couldn’t get any costlier, Nvidia and even AMD could find yourself climbing costs for each GPU out there now.

In the previous week, studies stemming from Taiwan-based PC {hardware} tipster BenchLife prompt Nvidia was proposing one other spherical of value hikes for its GPUs, each its “Founders Editions” made by Team Green itself and its add-in board (AIB) companions—the separate producers who construct the overwhelming majority of obtainable graphics playing cards. Nvidia sells its GPUs and VRAM (video random entry reminiscence) as an all-in-one package, which means the AIB corporations don’t have any alternative however to just accept increased prices, which naturally get handed to shoppers.

The information will get even worse. BenchLife prompt the prices would impression present GDDR7 reminiscence modules—utilized by the RTX 50-series—in addition to older GDDR6. While Nvidia reportedly ended manufacturing of its older RTX 40-series GPUs, the reminiscence costs could certainly have an effect on different graphics playing cards. AMD depends on GDDR6 reminiscence for its Radeon RX 90-series GPUs.

Rumors on-line swirled that AMD was planning value hikes for this month, however August could lastly see AMD’s GPU costs spike in tune with Nvidia’s pricey playing cards. Semi-reliable rumors stemming from WeChat tipster Board Channels (by way of VideoCardz) recommend AMD delayed value will increase till Nvidia made related changes to its AIB pricing. Gizmodo reached out to AMD for remark, and we’ll replace this submit if we hear again.

Compared to the RTX 50-series, AMD’s graphics playing cards just like the Radeon RX 9060 XT, RX 9070 and RX 9070XT have remained nearer to their 2025 value factors, thanks partly to utilizing the older sort of VRAM. The Radeon maker had been explicitly making an attempt to undercut Nvidia, although it too felt the reminiscence crunch sufficient to launch an AMD Radeon RX 9070 GRE within the U.S. with solely 12GB of VRAM as an alternative of 16GB on the opposite playing cards in its class.

AMD’s management beforehand promised they’d attempt to preserve GPU costs low, however there’s no method any of those corporations can stave off the ballooning value of reminiscence pushed by the push for AI hyperscalers. As all the time, the ever-present irony is these corporations are feeding the beast that retains reminiscence costs so excessive. Last week, AMD showcased its new datacenter-centric rack-scale chips, dubbed Helios. These are supposed to compete towards Nvidia’s Grace Blackwell and Vera Rubin AI chips.

The main graphics card makers have largely deserted any point out of their mid-gen GPU refreshes resulting from these spiking costs. Nvidia has stored the lid on its RTX 50-series “Super” graphics playing cards, and we could not see them till 2027.

The RTX 50-series, particularly playing cards sporting 16GB or extra VRAM, had been already astronomically costlier than they had been at launch. An Nvidia GeForce RTX 5070 Ti launched with an MSRP of $750. Now, you’ll be able to solely discover the cardboard for $1,100 or extra, and that’s together with gross sales costs. A comparable RX 9070XT calls for just below $800, which continues to be greater than its preliminary launch value.

Don’t even hassle glancing at top-end GPU costs. Nobody, and I imply no person, ought to truthfully think about shopping for an Nvidia GeForce RTX 5090 for $4,300 or extra, regardless of what number of frames you think about you’ll be able to push on Nvidia’s flagship.