Nvidia Reportedly Cancels RTX 5080 24GB Over GDDR7 Memory Costs
3GB GDDR7 chips now cost about three times as much as 2GB ones, so the January 2027 launch is reportedly off and the 16GB RTX 5080 stays the top pick.
Hakkı ALKANPublished 3 min read
Nvidia has reportedly canceled the January 2027 launch of the RTX 5080 24GB, a higher-memory version of its current high-end gaming card. According to a November 10 report from Chinese industry site Board Channels, Nvidia told its graphics card partners that the plan is off and gave no new date, because the 3GB GDDR7 memory chips the card needs cost about three times as much as the 2GB chips used today.
Nvidia has never officially confirmed that an RTX 5080 24GB existed, so treat this as a rumor. Still, if it holds, the regular RTX 5080 16GB stays the fastest GeForce most buyers can realistically get until the next generation arrives.
Why Was the RTX 5080 24GB Canceled?
It comes down to memory prices. The RTX 5080 uses eight GDDR7 memory chips on a 256-bit bus. Swapping each 2GB chip for a 3GB one would raise capacity by 50%, from 16GB to 24GB, without a new GPU or a wider bus. That makes it a cheap upgrade on paper.
In practice, the report says, 3GB GDDR7 modules now cost roughly three times as much as 2GB modules. Board Channels describes the move as a "temporary cancellation" made because the price of 3GB chips cannot come down for now. As VideoCardz notes, the same memory costs were already linked to delays for the RTX 50 SUPER refresh, which was also expected to bring a 24GB RTX 5080 SUPER.
What Changes for the RTX 5080 16GB?
The 24GB card was reportedly meant to replace the 16GB model as Nvidia's flagship GeForce in China. Earlier reports even said Nvidia would wind down shipments of the RTX 5080 16GB and RTX 5090 to make room for it. Now Nvidia will reportedly keep supplying GPU kits with 2GB chips, so partners can keep building the 16GB card.
This is also a sharp reversal. In late October, the same source claimed Nvidia was pulling the 24GB launch forward to December 2026 or January 2027.
| Model | Memory | Status |
|---|---|---|
| RTX 5080 16GB | 8 x 2GB GDDR7, 256-bit | On sale since January 2025 at a $999 MSRP |
| RTX 5080 24GB | 8 x 3GB GDDR7, 256-bit | Reportedly canceled for January 2027 |
| RTX 5080 SUPER | 24GB GDDR7 (expected) | Reportedly delayed |
How Much Would It Have Cost?
Earlier rumors, cited by Club386, put the RTX 5080 24GB at around $2,999. That is far above the RTX 5090's original $1,999 MSRP, for a card with 8GB less memory and less than half the CUDA cores. Besides memory, few changes were expected: the same core count, with possibly slightly higher clocks and power.
Graphics card prices are already far from normal. We recently reported that the RTX 5090 price hit $9,400 at some US retailers.
Memory is the common thread. The same AI-driven demand explains why RAM is so expensive right now.
Should You Buy an RTX 5080 Now or Wait?
For most gamers, 16GB of video memory (VRAM) is still enough for 4K gaming in current titles. Extra VRAM matters more if you run local AI models, heavy texture mods or path tracing at very high resolutions. There is also no sign of a SUPER refresh soon, so the next big step is likely the RTX 60 series based on Nvidia's Rubin architecture.
In our view, this rumor fits a pattern: memory makers are prioritizing AI data centers, and gaming cards are getting the leftovers. That is partly why AMD outsold Nvidia on Amazon US recently. AMD, meanwhile, is reportedly planning only one RDNA 5 desktop GPU for 2027.
If a relative asks what happened, the short answer is this: Nvidia reportedly shelved a 24GB RTX 5080 because the memory became too expensive, so the 16GB version stays on shelves.