RTX 5090 GPUs Reworked for AI Servers Amid Supply Shortage
Footage shared by Hong Kong hardware outlet HKEPC on 23 October shows batches of new Nvidia RTX 5090 graphics cards being dismantled and rebuilt as two-slot blower cards for AI servers. Workers remove the original coolers and transfer the GPUs and memory to new circuit boards, then test the finished cards. The thinner design exhausts heat out of the server and allows more cards to fit side by side.
The video does not confirm where the factory is or how the cards reached it. Reports have pointed to China, where the full RTX 5090 is barred from sale under US export controls; the article says the footage alone does not establish how the cards were sourced.
Nvidia executive Jeff Fisher denied on 16 October that the RTX 5090 had been discontinued, but acknowledged supply constraints without giving a timeline for improvement. US retailer listings cited in the article put many RTX 5090 cards above $9,300, compared with a $1,999 suggested retail price. The RTX 5090 conversion highlights strong AI-server demand and tight GPU supply, but its scale and effect on consumer availability remain unclear.
Neutral
The report has no direct, confirmed catalyst for cryptocurrency prices. RTX 5090 cards are being repurposed for AI servers, not identified as equipment for crypto mining, and the article gives no evidence of changes in blockchain activity, token demand, or crypto-market liquidity. The most direct effects are on GPU availability and pricing.
In the short term, traders may read the footage as further evidence of strong AI-computing demand and constrained high-end GPU supply. That could influence sentiment around AI-related equities or crypto projects marketed around decentralized computing, but any spillover into tokens would likely be indirect and sentiment-driven. The factory’s location, the cards’ source, and the scale of the operation remain unverified, limiting the strength of that signal.
Over the longer term, sustained competition between AI workloads and consumer demand could keep pressure on GPU prices and availability. Historical AI-hardware shortages have sometimes supported narratives around computing infrastructure, while hardware and supply-chain headlines have not reliably translated into broad crypto price moves. Traders should therefore treat this as a technology-supply story rather than a standalone bullish or bearish signal for digital assets, and look for confirmation in token-specific demand, trading volume, and broader risk-market conditions.