Although NVIDIA has denied reports that the RTX 5090 has been discontinued, rumors continue to circulate. Recently, a foreign hardware-leak YouTuber claimed that NVIDIA’s decision to cut high-end gaming cards has nothing to do with new products; instead, it is worried that the AI investment boom cannot last much longer, so it wants to use its limited chips to prioritize higher-margin professional and AI products while demand still exists—selling as many as it can.
Image source:VideoCardz
“Is the AI party almost over?” A foreign leaker claims NVIDIA may have discontinued the RTX 5090 and 5080, hoping to sell AI GPUs in bulk before the bubble bursts.
For readers who aren’t very familiar with the news that the RTX 5090 and RTX 5080 have been discontinued, here’s a quick explanation.
This began on October 10, when a Chinese leaker, hongxing2020, claimed to have received notice from the factory that NVIDIA will no longer produce any RTX 5090, including the China-exclusive D version, and that the RTX 5080 16GB will also be discontinued, with the RTX 5080 24GB taking over as the highest-end GeForce model.
5090 EOL🙀，5080 24GB ready 🚀 https://t.co/FUc1igiwlD
— hongxing2020 (@hongxing2020) October 10, 2026
On the same day, another leaker, MEGAsizeGPU, also said on X that the GB202 chip used in the RTX 5090 will later be exclusive to RTX PRO professional cards and will no longer be used to produce new GeForce RTX 5090s.
All future GB202 supplies are RTX Pro exclusive
— MEGAsizeGPU (@Zed__Wang) October 10, 2026
The overseas hardware leak YouTube channel MLID also cited a source working at a large retailer on its own show Broken Silicon, saying that the last batch of RTX 5090 stock arrived in September, while the last batch of RTX 5080 arrived by the end of October; the two graphics cards are expected to stop production in the first half of January 2027. A placeholder item for the RTX 5080 24GB has also appeared in the retailer’s system, expected to launch in early 2027 at a higher price than the 16GB version.
NVIDIA Senior Vice President of Personal Computing Jeff Fisher, in an interview with the media after the “RTX Spark Launch Party: Taipei” event, clearly stated that the RTX 5090 has not been discontinued and will continue to ship.
MLID disagrees with this statement.
According to foreign media Notebookcheck According to reports, MLID said in a video a few days ago that end-of-sale dates for the RTX 5090 and RTX 5080 16GB have already appeared in retailer systems. Moreover, the RTX PRO 5500’s specs are almost identical to the RTX 5090’s, and it was announced one month earlier than the news of the RTX 5090’s discontinuation emerged, meaning NVIDIA has begun diverting its already limited consumer-grade GB202 supply to the professional market.
He speculates that NVIDIA seems worried the AI boom won’t last long, so it is trying to ship as much as possible while customers are still buying heavily, thereby securing higher profits:
In addition, MLID specifically mentioned the “AI debt bomb,” claiming that reports have already indicated that NVIDIA customers have warehouses piled with large numbers of GPUs that have not been brought online, and believes that CoreWeave may not be able to hold out before its data centers are completed. Once the AI bubble bursts within 12 months, data centers that are not yet completed will also not be able to serve as valuable collateral:
Of course, all of this is just MLID’s speculation. For general hardware enthusiasts, they might hope the AI boom cools down so that prices of components such as graphics cards and memory gradually return to normal, but the current evidence at most shows tight supply of high-end GeForce and possible changes in GB202 chip capacity allocation; it is not enough to prove that the AI bubble is about to burst. As for whether the RTX 5090 and RTX 5080 16GB will really be discontinued in January 2027, we will know in just over two months.
Source: KOCPC Chinese