# RAM Is Just The Beginning: Why The 2026 Tech Apocalypse Will Be Different…

Source: https://www.youtube.com/watch?v=KZeaY_mnAVI
Recap page: https://rapidrecap.app/video/KZeaY_mnAVI
Generated: 2025-12-30T18:33:46.021+00:00

---
## Quick Overview

The upcoming tech apocalypse driven by AI demand for memory and GPUs will be different from past cycles because memory manufacturers are deliberately constraining supply to maintain profitability until 2028, leading to perpetually inflated consumer GPU prices despite cooling crypto demand.

**Key Points:**
- The GPU market cycle is predicted to enter a new cycle (a 'bust') because AI demand is now consuming memory and storage supply at a massive scale, setting the stage for a decade-long pricing issue.
- Unlike the 2017 crypto boom, the current situation involves enterprise AI data centers (like those of Google and Microsoft) aggressively buying up memory, which is the primary driver of current shortages.
- Memory manufacturers Samsung and SK Hynix are signaling they will actively minimize the risk of oversupply by constraining production until possibly 2028, directly impacting consumer GPU availability and pricing.
- The speaker specifically notes that in 2017, he missed out on an EVGA RTX 3080 at MSRP because of the crypto mining rush, which sold out immediately.
- The current situation is worse because the enterprise AI demand is so large that it swamps consumer demand; for example, Microsoft acquired nearly 500,000 NVIDIA 'Hopper' GPUs for AI that year.
- The speaker forecasts that consumer GPU prices will likely remain high or even increase due to this massive enterprise demand, unlike the previous GPU busts where prices eventually fell back to MSRP.
- The second major difference is that AI demand is so high that it creates a constant, non-cyclical demand floor, unlike the volatile crypto market of the past.

![Screenshot at 00:00: The presenter introduces the video by drawing a rough 'GPU Market Cycle' graph on a whiteboard, illustrating past peaks in 2016/2017 and 2020/2021, followed by a question mark indicating uncertainty for the current/future cycle.](https://ss.rapidrecap.app/screens/KZeaY_mnAVI/00-00-00.jpg)

**Context:** The video analyzes the current and future state of the GPU market, drawing parallels and highlighting differences with the 2017 GPU shortage caused primarily by cryptocurrency mining. The speaker uses historical price charts, news articles from Tom's Hardware and PC Gamer, and an article about the AI arms race to argue that the current demand from AI data centers is fundamentally different and more enduring than previous boom cycles, suggesting that consumers face a prolonged period of high GPU prices, potentially lasting until 2028.

## Detailed Analysis

The presenter argues that the current GPU shortage and high pricing environment is fundamentally different from the 2017 crypto boom due to the massive, sustained demand from enterprise AI data centers, which are consuming memory and storage supply at an unprecedented scale. He cites an article stating that AI data centers are 'swallowing the world's memory and storage supply, setting the stage for a pricing apocalypse that could last a decade' (08:13). This AI demand is inelastic and persistent, unlike the volatile crypto demand. Furthermore, memory manufacturers like Samsung and SK Hynix are explicitly signaling a strategy to maintain long-term profitability by minimizing the risk of oversupply, potentially keeping constraints in place until 2028 (10:51). This commitment means that even if crypto mining demand collapses, consumer GPU prices will not return to MSRP quickly, as AI demand continues to absorb supply. The presenter notes that in the 2017 cycle, gamers benefited from the eventual bust when miners offloaded used cards, but this time, the massive enterprise demand keeps the market tight, suggesting that consumers will continue to pay inflated prices for the foreseeable future. He concludes that the current situation is worse because the demand floor set by AI giants like Google and Microsoft is far more stable and massive than the crypto demand ever was.

### GPU Market Cycles Analysis

- The 2017 cycle was driven by crypto booms; the 2020 cycle involved both crypto and the pandemic; the current cycle (starting 2023/2024) is driven by AI data centers
- Past cycles showed significant price drops post-bust, but the current AI demand creates a higher price floor. (00:03, 01:01, 04:21, 07:28)

### 2017 Crypto Boom Impact

- The crypto boom caused GPU prices to soar, leading to shortages where the presenter himself could not buy an RTX 2070 at MSRP in 2017 (2:00, 2:11, 2:27).

### The AI Demand Shift

- Enterprise AI is now the primary driver, with companies like Microsoft acquiring nearly 500,000 NVIDIA 'Hopper' GPUs (09:25). Google is projected to spend trillions on data centers (09:50).

### Memory Supply Constraint

- Samsung and SK Hynix are deliberately constraining DRAM production until potentially 2028 to avoid oversupply and maintain profitability, which will keep GPU prices high (10:41, 10:51).

### Future Outlook & Comparison

- The current situation is worse because AI demand is less volatile than crypto; the presenter estimates a 100% chance of a GPU price drop in the near term but warns that the underlying demand structure will keep prices inflated long-term (11:34, 12:29, 12:53).

![Screenshot at 00:07: Presenter introduces the concept of GPU market cycles using a hand-drawn graph on a whiteboard.](https://ss.rapidrecap.app/screens/KZeaY_mnAVI/00-00-07.jpg)
![Screenshot at 00:10: Text overlay highlights the video's focus on GPU shortages over the last 10 years.](https://ss.rapidrecap.app/screens/KZeaY_mnAVI/00-00-10.jpg)
![Screenshot at 01:15: Meme showing empty shelves labeled 'GeForce GTX' and 'Radeon' with the text 'The Great GPU Famine Never forget 2017'.](https://ss.rapidrecap.app/screens/KZeaY_mnAVI/00-01-15.jpg)
![Screenshot at 02:28: CoinMarketCap chart showing the massive Bitcoin price surge during 2017, correlating with the first GPU shortage.](https://ss.rapidrecap.app/screens/KZeaY_mnAVI/00-02-28.jpg)
![Screenshot at 09:21: News headline from TechPowerUp stating 'Microsoft Acquired Nearly 500,000 NVIDIA "Hopper" GPUs This Year', illustrating massive AI demand.](https://ss.rapidrecap.app/screens/KZeaY_mnAVI/00-09-21.jpg)
