Nvidia is reportedly trying to reduce gaming GPU manufacturing by as much as 40% in 2026 as a consequence of VRAM provide points, but it surely's not as unhealthy information as you may suppose. Not but, no less than

It’d really feel like tech information is a caught report for the time being, with each different headline reporting on the grim results of the present DRAM provide disaster. And at first look, a report that Nvidia is planning on chopping discrete GPU manufacturing by as much as 40% simply looks like a living proof. Nonetheless, whereas it’s all RAM-related, it is not essentially unhealthy tidings.

The declare in query was posted on the Chinese language Board Channels discussion board (through Benchlife and Overclock3D) and principally claims that Nvidia is ready to cut back the availability of its gaming GPUs by between 30% and 40% within the first half of 2026, in comparison with the quantity of output throughout the identical interval in 2025.

That is not stunning on condition that the DRAM provide issues, affecting the whole computing trade, will lead to fewer desktop and laptop computer PCs being shipped subsequent yr. Given simply what number of of them sometimes sport an Nvidia discrete GPU, both soldered onto the motherboard or within the type of a third-party graphics card, the demand for RTX Blackwell chips is clearly going to plummet.

Nonetheless, there’s not precisely an enormous scarcity of GeForce RTX 50-series playing cards for the time being, except sensibly priced RTX 5080 and 5090 fashions. Flick thru any retailer’s listings, and you will see countless provides of RTX 5050, 5060, and 5070 graphics playing cards. This actually wasn’t the case 10 months in the past, however Nvidia and its companions ramped up manufacturing significantly because the launch of Blackwell.

So that you’re unlikely to see RTX 50-series playing cards all of a sudden disappearing from cabinets, although the report additionally means that some add-in board distributors plan to regulate provides of 16 GB fashions first, specifically the RTX 5060 Ti and RTX 5070 Ti playing cards. Once more, that is sensible as a result of these use twice as many GDDR7 modules because the 8 GB RTX 5060 Ti and 33% extra chips than the RTX 5070.

Nvidia RTX 5070 Ti graphics card

RTX 5070 Ti playing cards have the identical variety of VRAM chips as RTX 5080 playing cards, however make Nvidia much less cash. (Picture credit score: Future)

There’s one other issue to think about in all of this, which is not coated within the report: Nvidia’s anticipated RTX 50 Tremendous variants. The overall assumption for these playing cards, based mostly on the quite a few leaks and rumours which have cropped up all through 2025, is that these will use 3 GB GDDR7 modules as an alternative of the usual 2 GB ones.

RTX 5090 laptops already use these, which is why they sport 24 GB of VRAM, despite the fact that the GPU inside solely has a 256-bit reminiscence bus (256 bits comes from 8 x 32-bit broad reminiscence buses, so 8 x 3 GB = 24 GB). Ought to an RTX 5080 Tremendous, RTX 5070 Tremendous, and presumably even an RTX 5060 Tremendous all sport these measurement VRAM chips, then Nvidia must deal with the truth that whereas it has sufficient GPUs to fulfill the demand for these playing cards, it may possibly’t assure the identical for the VRAM.

The answer? Lower the availability of ‘regular’ GDDR7-equipped chips in order that the VRAM suppliers are in a position to produce adequate reminiscence chips for the Tremendous fashions. So far as I can inform, solely Samsung is producing 3 GB GDDR7 modules to any diploma of amount, and if Nvidia is already utilizing the corporate to produce 2 GB chips, swapping the order books over to the upper density modules means much less VRAM goes to be obtainable for its common Blackwell chips.

That stated, there was a hearsay that Nvidia can also be planning to cease packaging VRAM modules with its GPUs (Asus, MSI, and so forth purchase each of them from Nvidia, as issues at present stand), so if that involves cross, then possibly the information is grim studying.

A close-up photo of an Nvidia RTX 4070, with its heatsink removed, showing the AD104 GPU die and the surrounding Micron GDDR6X VRAM chips

Cannot have a graphics card with out VRAM chip, sadly. (Picture credit score: Future)

Since RTX Blackwell is not even a yr outdated but, the following era of GeForce GPUs is not going to look any time quickly, so Nvidia will need to have one thing new for the graphics card market in 2026. That may solely be RTX 50 Supers, and if these contemporary fashions have already been deliberate and lined up for launch, Nvidia might want to do every little thing it may possibly to make sure that they’re truly available for purchase.

Thoughts you, there wasn’t a RAM provide disaster when Blackwell launched earlier this yr, and but provides have been so unhealthy that the majority GeForce RTX 50 playing cards have been both extremely laborious to search out or they sported a price ticket a whole bunch of {dollars} over the MSRP. I do not drink alcohol, however I am nonetheless taking place to the Winchester to have a pint of one thing fizzy and wait till this all blows over.

Greatest graphics card 2025

Supply hyperlink