NVIDIA

You Probably Won't Be Buying A GeForce RTX 5090 In January, But What About The 5080?

At this point, it would be absolutely shocking of NVIDIA did not go forward with rumored plans to unveil its initial next-generation GeForce RTX 50 series lineup at the Consumer Electronics Show (CES) next month. NVIDIA CEO Jensen Huang is delivering an open keynote at the event, and the recent GeForce LAN 50 announcement are strong indicators that a formal launch is imminent. But when will you actually be able to buy a next-gen GPU?

That's the bigger question, and the answer may depend on which GeForce RTX 50 model you're interested in purchasing. It's widely expected that NVIDIA will unveil its next flagship GPU, the GeForce RTX 5090, along with the GeForce RTX 5080, GeForce RTX 5070, and possibly a GeForce RTX 5070 Ti at the event. A mobile GPU launch could happen as well, based on recent leaks and rumors.

It's not unusual for NVIDIA to start at the high-end of the GPU spectrum with new generation product launches. As it pertains to the GeForce RTX 50 series, the folks at WCCFTech say they are privy to "information regarding the actual retail launch" of the first batch of SKUs.

Assuming their information is accurate, the GeForce RTX 5080 will be the first one to land at retail, with availability expected around the middle of January. Based on past leaks/rumors, the GeForce RTX 5080 is expected to wield 10,752 CUDA cores and 16GB of GDDR7 memory (32Gbps) tied to a 256-bit bus, for 1,024GB/s of memory bandwidth. It's also said to be a 400W part.

The GeForce RTX 5090, meanwhile, will hit retail shelves a little while later, though fortunately (for interested parties) not a whole lot longer—it's said to be landing in late January or by the middle of February. Both cards are rumored to feature enhanced ray tracing and neural rendering tech.

By all accounts, the GeForce RTX 5090 is going to be a monstrous GPU. Based on NVIDIA's GB202-300 Blackwell GPU, it's expected to arrive with 21,760 CUDA cores underneath the hood, as well as a heaping 32GB pile of GDDR7 memory (28Gbps) linked to a fat 512-bit bus for 1,792GB/s of memory bandwidth.

It's an interesting strategy on multiple levels. For one, as previously rumored, it's possible that the GeForce RTX 5080 will employ faster memory chips than the GeForce RTX 5090 (32Gbps versus 28Gbps). That won't affect the latter's flagship status, as it will have twice as much memory and a wider memory bus to boot.

The other interesting aspect here is the supposed decision to make the GeForce RTX 5080 available first, if that indeed occurs. It's the opposite strategy NVIDIA employed with the GeForce RTX 40 series, which saw the GeForce RTX 4090 reach retail weeks before the GeForce RTX 4080.

It is anyone's guess why NVIDIA might reverse the order next round. One possible reason is that a potential price increase for $1,999, which could make the next-gen flagship even more of a comparatively niche SKU. Related, NVIDIA could be hoping to spur a flurry of initial sales, and a less expensive GeForce RTX 5080 (compared to the 5090) could accomplish that goal.

mundophone