Same RAM. Nearly six times the price.

I’ve been looking at what it would take to build a really good research workstation. A Threadripper, plenty of RAM, maybe one of the Blackwell workstation GPUs. Definitely getting ahead of myself, but it’s fun to put the thing together in your head.

Then you start pricing it out. Holy shit.

Somewhere in all of this, we found a 32 GB Corsair DDR5 kit listed at $484.99 on Newegg. The same kit had been recorded at $81.99 in July 2025. We checked it on September 4, 2026. Nearly six times the price in a little over a year.

You can get other kits for less, but I wanted to know why this one had gone up so much.

The more we looked into hardware prices, the more I wanted to know. What did these things cost a few years ago? Is there actually a better month to buy? Could we see anything interesting if we compared prices with the weather? That last one was probably unnecessary, but those are the kinds of questions I want to be able to ask just to see what happens.

That’s also how a workstation shopping trip ends up with a whole page of charts and no new workstation.

My first thought about the prices was probably the same as yours: AI. And there is something to that. Memory manufacturers have been shifting production toward servers and AI demand, leaving less available for PC memory. That doesn’t explain every dollar on this particular listing, but it helps explain why a desktop upgrade can get caught up in something much bigger.

I find that frustrating because I’m genuinely excited about where this is all going. I want to run models locally, try different experiments, and learn enough to build things I wouldn’t have known how to start a year ago. Then I look at what the equipment costs and think, okay, maybe we need to slow down a little.

I really enjoyed digging into the history, though. I’d look at a chart and immediately want to see what happened if we went further back or looked at another piece of hardware. There was always something else I wanted to check.

I have a growing list of things I want to build and experiments I want to run. Looking into the hardware is part of that, but I’m much more interested in what we could actually do with it. I’m curious what other people would choose to work on if they had access to the setup they wanted.

I’d like to see more people doing research because something caught their attention and they wanted an answer. Or building something simply because they want it to exist. There’s a lot to be excited about there, even if the hardware side is frustrating right now. I hope it gets easier to get the equipment too. I’d rather be comparing what we can do with it than what we can afford.

For now, I’m keeping the current setup and putting it to work. There’s more I want to do with it, including making better use of the CPU. I’m still looking at parts for the next workstation, though. That build is going to be fun when the time comes.

I put the charts and the data behind them here if you want to have a look. The seasonal comparisons are in there too.