No points for guessing this one in advance. The morning of the official retail launch for Nvidia’s next-gen GeForce RTX 50-series graphics cards, they’re almost impossible to actually order at online stores in the United States just an hour after the sales began. Since lines were seen forming days ago at brick-and-mortar retailers, it’s safe to assume a similar situation is in play there.
Despite prices mostly elevated from the initially announced MSRP of the RTX 5080 and RTX 5090, every single one of the 47 different variations on Newegg’s site is now sold out, from the $999.99 base model RTX 5080 to the $2,799.99 commanded for the Asus ROG Astral RTX 5090. Some of these may or may not have actually been available for purchase, as many of the cards listed on Best Buy (also completely sold out) are shown as “coming soon.”
It’s perhaps entirely expected that the new cards would be difficult to find at launch, as that’s been the case for Nvidia launches for years. And since it’s been two years since a significant generational jump for RTX, this is hardly a surprise. Even so, Nvidia seems to have particularly low supplies this time around, something it warned of in advance. Speaking with retailers and third-party manufacturers, we’ve been told to expect the low supply to continue. MSI, one of the more prolific Nvidia partners, doesn’t expect to have RTX 5090 stock to ship from in its online store until February 6.
Are these cards worth the price, or indeed, the multiple hundreds of dollars more that some of the gently-upgraded AIB designs demand? That depends on what you’re looking for. While the full-fat RTX 5090 ($2000 for Nvidia’s self-branded Founder’s Edition) seems to cut the mustard as the new most powerful card on the market, the RTX 5080 has disappointed in initial reviews, offering a much more modest power upgrade over the last generation and leaning hard on Nvidia’s DLSS and frame generation technologies.
Editorially, I wonder if these initial supply issues have anything to do with the still-booming AI market. Massive AI data centers are mostly powered by discrete GPUs, most of which are Nvidia chips. While a lot of these are more bespoke B2B designs, it’s more than possible to build up that “big iron” machines on the same cards coveted by PC gamers.
I’d also beg our readers not to enrich the scalpers who are instantly flipping the latest GPUs for double or triple profit. If you don’t stop paying them, they won’t stop doing it. However small the supply of $1000+ graphics cards, there are only so many people with the kind of disposable income it takes to afford them, and you’ll almost certainly be able to find a GPU at retail price within a few months.
