by 
Earlier this month we noted that power supply maker Seasonic had added entries for three RTX 50-series Super graphics cards to its system power calculator, showing increased TDPs and some level of anticipation for the new cards. Now, Videocardz is reporting that the Super cards have reached at least one add-in board partner (eg the likes of Asus, Gigabyte and MSI) - but that the launch of the GPUs has been delayed.
The report states that this is due to the high cost of 3GB GDDR7 memory modules, which all three new Super cards use in order to offer higher VRAM counts than their vanilla counterparts that shipped with 2GB modules. Those changes see the RTX 5070 tier move from 12GB to 18GB, while the RTX 5070 Ti and RTX 5080 cards move from 16GB to 24GB with the Supers.
According to the report, Nvidia and its partners are facing the prospect of paying around $60 to $70 for a 3GB memory module, versus just $20 for a more broadly available 2GB counterpart. With six memory modules on the 5070 Super and eight on the 5070 Ti Super and 5080 Super, that accounts to a $270 price rise for the former and a $360 price increase for the latter, if we assume a middle-ground $65 cost.
Given that these are already expensive products - $630 for an RTX 5070, $920 for an RTX 5070 Ti and $1250 for an RTX 5080 - it's understandable that Nvidia would put its plans on hold. If we add on the extra cost for 3GB modules, ignoring any other additional costs, that would see the "mid-range" RTX 5070 Super at $900, the "upper-mid-range" 5070 Ti Super at $1280 and the "high-end" 5080 Super at a scarcely believable $1610 (!!). For context, the RTX 4090 launched at $1500 in 2022 with 24GB of GDDR6X.
The firms are therefore delaying the launch until memory prices improve - which could take quite some time given the general state of the PC industry as AI companies continue to monopolise the limited supply of flash memory components.
What do you make of this prospective pricing? Are you eager to upgrade to a new RTX Super card, given the extra VRAM allocation, or would you instead stick with the existing RTX and Radeon options? Let me know in the comments below.