Massive Tech’s AI-fueled reminiscence scarcity is about to be the PC business’s defining story for 2026 and past. Standalone, direct-to-consumer RAM kits had been among the first merchandise to really feel the chew, with costs spiking by 300 or 400 % by the top of 2025; costs for SSDs had additionally elevated noticeably, albeit extra modestly.
The remainder of 2026 goes to be all about the place, how, and to what extent these value spikes move downstream into computer systems, telephones, and different parts that use RAM and NAND chips—areas the place the prevailing provide of merchandise and longer-term provide contracts negotiated by large corporations have helped hold costs from surging too noticeably to this point.
This week, we’re seeing indicators that the RAM crunch is beginning to have an effect on the GPU market—Asus made some waves when it inadvertently introduced that it was discontinuing its GeForce RTX 5070 Ti.
Although the corporate has since tried to stroll this announcement again, in the event you’re a GPU producer, there’s a powerful argument for both discontinuing this mannequin or de-prioritizing it in favor of different GPUs. The 5070 Ti makes use of 16GB of GDDR7, plus {a partially} disabled model of Nvidia’s GB203 GPU silicon. This is identical chip and the identical quantity of RAM used within the higher-end RTX 5080—the considering goes, why proceed to construct a graphics card with an MSRP of $749 when the identical primary elements may go to a card with a $999 MSRP as a substitute?
Whether or not Asus or some other firm is canceling manufacturing or not, you may see why GPU makers can be tempted by the argument: Avenue costs for the RTX 5070 Ti fashions begin within the $1,050 to $1,100 vary on Newegg proper now, the place RTX 5080 playing cards begin within the $1,500 to $1,600 vary. Although 5080 fashions may have extra sturdy boards, heatsinks, and different parts than a 5070 Ti, in the event you’re simply attempting to maximise the profit-per-GPU you will get for a similar quantity of RAM, it is smart to shift allocation to the costlier playing cards.
