It’s no secret that GPU prices have been driven up by ongoing memory shortages, with high-end models the worst affected. Despite Nvidia’s flagship RTX 5090 now costing more than $6,500 on Amazon, we’ve noticed multiple models running low on stock (some with only one available at the time of writing). Recent reports suggest AI firms are buying them in bulk.

Nvidia’s data center-focused RTX PRO 6000 card with 96GB jumped to $16,000 (more than double its original pre-order price) last month, and reports reveal AI firms are now turning to the flagship gaming model, which comes fitted with 32GB of GDDR7 VRAM for reference. Elsewhere, we’ve seen stores selling the RTX 5090 for as much as $9,800 with overseas shipping.

AI firms are reportedly buying the RTX 5090 in bulk

Hong Kong-based publication HKEPC reports that “many Chinese vendors” [translated] are stocking up on RTX 5090 from various retailers for use in AI workstations, which are then to be sold on. The outlet shared multiple images showing pallets full of RTX 5090 cards, many of which are cards from ASUS, MSI, and Gigabyte, typically marketed towards gamers.

There’s an AI disclaimer on these images, and on closer inspection – particularly around model names/numbers – some AI image generation is going on here, so be aware of that. Interestingly, many of the cards look legitimate, which seems to suggest these images have been edited to showcase more stacks of GPU boxes than there are in reality, for whatever reason.

It’s worth pointing out that the cards pictured do not appear to be the China-specific RTX 5090D or the RTX 5090D v2 (with less VRAM), but the same RTX 5090 cards we see sold internationally. The ‘D’ models have their AI capabilities toned down to comply with export restrictions, maxing out at 2,375 AI TOPS as opposed to the original RTX 5090’s 3,352 AI TOPS.

About the Author

Jack Goodall

At PC Guide, Jack is mostly responsible for reporting on hardware deals. He also specializes in monitors, TVs, and headsets and can be found putting his findings together in a review or best-of guide.