If you like to build your own PCs, upgrade what you have, or tinker in any way whatsoever, you have likely run into inflated costs as of late. Gone are the days of $600 flagship video cards and affordable RAM; now, affordable NVMe (SSD) solutions are no longer available, as these products join the party.
Video cards (GPUs) have been notoriously expensive ever since cryptocurrency mining became popular. Pricing has followed global trends ever since, with companies even shifting focus to special versions of their GPU products just for these trends (ie, video cards specially designed around mining instead of gaming and other tasks).
As this was happening, data centers were ramping up construction and deploying larger build-outs to accommodate the cloud data industry. From much larger data packages to intense cloud-driven operations in both gaming and other GPU-intensive operations, these data centers required better GPUs and were ordering in bulk. Thus, GPU companies started shifting focus to these demands since these demands led to stronger profits. Causing consumers to suffer from sudden price hikes caused by all of this. Why not make the innocent suffer and be charged with the R&D behind everything else, instead of shifting this punishment to the industries causing the issues to begin with, right? I know, it’s a bit of a loaded question.
Through all of this, RAM prices and availability also began to suffer heavily. RAM prices could suddenly be tracked in a chart as if they were being traded on Wall Street. Up and down those prices went, while usually increasing in cost, versus any decreases that might offer consumers a little relief.
Then came the pandemic. This brought a whole new collection of excuses (ie, global supply chain shortages, increased operating costs, etc) to increase prices all over. In some situations, these excuses were valid, while others were just a fib to inflate profits. Some companies saw historical yearly profits, which didn’t exactly coincide with their claims of having to increase prices in order to survive.
Just to make things worse (as if it wasn’t bad enough), the AI industry has been exploding over the last few years, leading to a whole new demand for data centers. Suddenly, data centers needed a huge increase in power, as well as water to help keep things cool. Straining resources and raising utility costs for millions of people all over the country (or billions if you consider globally). And, of course, this also created a demand for storage as everyone flooded the gates to use these flashy new AI tools.
So now we make it toward the end of 2025, and RAM has been shooting right back up again, hitting painful numbers. But that wasn’t it, as suddenly, NVMe prices started to climb. In some cases, doubling in price for some of your favorite options. Take, for example, the Samsung 990 Pro 1TB NVMe. This drive used to hover around $119, sometimes seeing sales as low as $80 or less during the holidays. That same drive is now at an average of $199 or more, depending on where you shop. Some users are trying to hock these for closer to $300 in hopes that they can take advantage of the latest shortage.
NVMe costs are not expected to return to normal until 2027-2028 as companies continue to shift wafer capacity away from consumer-grade NVMe products and toward higher-margin enterprise products, prioritizing profitability over volume.
These companies could become peace-driven, for the people, ramping up production so that they could supply the needs of the many instead of isolating production toward the needs of big corporations. But they choose profit over the people, since their actions force prices to increase, which leads to higher profit margins for them. Proving that at the end of the day, no matter how innovative and exciting a company sounds when it comes to its passion to technologically evolve humanity into a brighter future, nothing matters more than numbers at the end of the day. These same companies would stop all innovative efforts if they felt it could maximize their profits with current and older tech.
In the short-run, this may lead to increased profits and happy shareholders, but it also breeds competition and R&D for alternative solutions, which could eventually undermine their leadership in the industry. One could hope, at least. Or, the AI bubble could completely burst, and prices could come collapsing down as companies scramble to stabilize themselves. I could continue to come up with additional scenarios, but these are all hopes, at best.
This leads us to one final thought — overexpansion. This is where I play a bit of devil’s advocate. If a company does ramp up production and the bubble ends up bursting (aka, a bad bet), this could lead to heavy losses as the company has to suddenly deal with the potential of excess inventory, layoffs, and more. So, suddenly ramping up production may not always be the solution. But that doesn’t mean solutions do not exist. It’s just that the last decade or so simply proved that companies couldn’t care less about finding these potential solutions.
So if you were “hoping’ that prices would come down anytime soon, so that you could affordably build that next computer or upgrade the one you have, you might be in for some disappointment. It will likely be a while until prices stabilize, and that’s assuming they ever will.
