Nvidia CEO Jensen Huang says its older GPUs are aging like “good wine,” but not in the way you’d hope
Table of Contents
Graphics card prices have been climbing for a while now, and despite some nice deals here and there, we’re still not at pre-‘RAMpocalypse’ levels. The growth of artificial intelligence has created a huge demand for powerful hardware, including GPUs. Because of this, even older graphics cards that launched years ago are suddenly becoming more expensive instead of cheaper.
According to Jensen Huang, older GPUs are now aging like “good wine.” The comment was made while discussing how strong the demand for AI hardware has become across the world. We tend to see terms like ‘fine wine’ used to describe GPU performance improvements over time, thanks to newer graphics drivers, not when prices go up.
Buying a GPU is like “investing in art” according to Jensen Huang
Older GPUs usually lose value over time as newer models arrive. Gamers often wait for prices to drop after a few years. But the AI boom has completely changed how the market works. Companies building AI models need massive amounts of GPU power, and many are willing to buy older hardware if newer chips are difficult to get.
Latest tech deals.
- AMD Ryzen™ 9 9900X 12-Core - was $499 now $328
- CORSAIR 4000D PC Case - was $124.99 now $75.99
- Hisense 65" Class E6SR - was $549 now $377.99
- Apple 2026 MacBook Air - was $1,999 now $1,799
- Skytech Gaming PC Desktop (9850X3d) - was $4,299 now $3,799
- Apple Magic Mouse - was $79 now $59
- MSI Prestige 16 Flip - was $2,149 now $1,899
- Kotin 9700X Gaming PC - was $2,399 now $2,199
Prices correct as of October 9th, 2026.
The term “good wine” has been used in the GPU industry before, especially during the AMD Vega era. Back then, it referred to graphics cards improving through better drivers and software updates over time. Nvidia is now using the same phrase differently. Instead of performance aging well, it is the prices that are going up. From consumer-friendly to not-so consumer-friendly.
Many AI data centers still rely heavily on older Nvidia AI hardware like the Nvidia H100, Nvidia H200, Nvidia A100, and Nvidia L40S. Even though some of these chips are already several years old, they remain extremely useful for AI training and cloud computing workloads. Of course, these are a completely different market from regular consumer GPUs.
One of the biggest reasons behind this situation is the supply shortage. Semiconductor companies are struggling to keep up with demand. AI companies, cloud providers, and tech firms are all trying to secure GPU inventory at the same time. As a result, prices across the entire hardware market are increasing.
This does not only affect enterprise hardware either. Consumer GPUs are also becoming harder to afford in some regions. Gamers who simply want to upgrade their PCs are now competing with AI companies for hardware availability. In some cases, even used GPUs are selling for much higher prices than expected.