The PC of Your Dreams Might be Hiding in a Company’s Surplus Bin

A version of this post originally appeared on Tedium, a twice-weekly newsletter that hunts for the end of the long tail.

I guess the saying is that one man’s trash is another man’s treasure. If that’s the case, I’ve been treasuring a lot of trash lately.

Between the $1 university I recently bought and the $10 Mac Mini I got working, I’ve been on a roll of maximizing the value of stuff that has long seen better days. And it’s with that in mind that I recently tested the waters of the upgrade arbitrage market, in an effort to build a workable home computer that I could use when I wanted a little more graphical oomph or I wanted to mess with some experiments.

Which led me into the fairly hefty arms of a Hewlett-Packard Z420 workstation that I got for $50.

Seven years ago, this machine could be had for just below $3,000 in its base configuration (and upgraded for many thousands of dollars more.)

My model, produced in 2013, supports many of the same components as the just-retired 2013 Mac Pro, including the same kinds of RAM and Xeon processors. Back when it first came out, this machine was probably used at an engineer’s desk, in a video-editing rig, or by a classroom of college students learning something suitably advanced and technical. I got a particularly low price because I bought local; you can find ’em on eBay for maybe $100-$150 with a processor.

It’s a great example of what I’ll call “upgrade arbitrage” in action.

The basic concept behind upgrade arbitrage, and why it’s so effective

If you’ve ever walked into a store like Marshalls, TJ Maxx, or Ross, you probably are familiar with the concept of arbitrage—the idea that you can exploit weaknesses in an economic structure to come out ahead.

In the case of these stores, you’re getting access to products that would cost much more at another store, but have issues that led to that product appearing in another market. Maybe the size is uncommon. Maybe there was a production overrun, or it was produced during the wrong season. Or maybe there’s a quality issue that’s imperceptible to you but was enough to prevent that shirt from going to Neiman Marcus.

Either way, you saved a lot of money by going to the cheaper store that focuses less on aesthetics and customer service and more on turning over inventory.

Arbitrage drives entire business models—for example, the recent trend of companies that literally sell bulk returns to people to resell on eBay at a higher markup. That guy who gives away money in all of his YouTube videos is likewise taking advantage of a similar disparity—he can make a ton of money through brand deals and ad revenue, and he can make more of it by spending some of that money in the most aggressively viral way possible.

More generally, if you’ve ever visited another country and benefited from a favorable exchange rate, you’ve practiced arbitrage.

Arbitrage works similarly when it comes to electronics or computing. Often a product no longer makes sense for its original target audience, so it goes on eBay or a surplus outlet, where it reaches someone for whom it does.

But in the case of what I’ll call “upgrade arbitrage,” the situation highlights a dramatic disparity between two markets. To put it simply: Back when a computer or similar product was originally released, it was prohibitively expensive for many consumers, while also being significantly more powerful than anything the average consumer could buy, anyway.

At the same time, there’s a general tendency for business-focused technology to trickle down to consumers after a while. Not everything—of course, there’s always that errant business-world feature that makes no sense in a consumer context. But give it a few years, and technology that’s just as fast will catch up to that big, bad machine from a few years prior.

That happens to be the point, surprisingly enough, where the business-focused gear depreciates enough in value that it becomes accessible to consumers who otherwise might not be willing to pay quite so much for it—but given its lower price, it’s suddenly significantly more valuable and can compete with modern, mainline hardware.

If a six-year-old workstation can get you within shouting distance of a modern consumer-grade computer for a lower price, even with a little not-made-for-the-home weirdness, you’re coming out ahead while wringing additional value from something that otherwise wouldn’t be worth very much.

This concept isn’t new. Far from it. Surplus auctions have been around forever, often in the context of military or higher education, and quite often, these auctions aren’t always heavily promoted to the public. And sometimes, it allows people to get a real deal on a machine that is no longer particularly valuable to its original owner.

To give you an idea, I found an interesting story when I was looking around for stuff for this piece. In 1969, a group of four students at the University of South Carolina did something pretty wild: They bought a gigantic IBM mainframe computer. The IBM 704 computer, when it was originally produced, cost $50,000 around 1961 ($427,300 when adjusted for inflation), when it was first sold to the U.S. Navy Mine Defense Laboratory in Panama City, Florida. (So few of them were sold that every installation was documented.) But around 1968, the laboratory started to upgrade its old gear, so they sold the computer at auction to this group of students.

“A lot of computer people around here are laughing their heads off at us,” one of the four buyers, C. Brian Honess, told the Associated Press. “But there are a few who think that maybe we’ll make a million dollars on the thing. All I know is we have what we wanted and we all think it will be fun to put this thing in operation. We think we got ourselves a bargain.”

The cost? $2,200 (or $15,300 today, a 97.4 percent discount when accounting for the inflation between when it was sold and when they bought it)—and they had to pick it up and attempt to rebuild it themselves. The goal? To get it working, give it a place to live (that was supplied by USC), and learn something along the way.

Honess later revealed what happened to this machine in an article he wrote for BYTE magazine in the late 1970s. Honess spent nearly 30 years teaching at USC before his passing in 2015, and occasionally moonlighted as a tech journalist. Here’s what his bio said:

> He reports that he has been an active ‘building’ radio amateur (ham) for 20 years; his interest in computers goes back to programming scientific business applications on an IBM 1620. He learned about what was Inside computers by buying a surplus IBM 704 from the government, and slowly taking it apart (donating, selling and scrapping the parts as he went). Another 704 was eventually purchased, and this time it was built back up, from the inside out. This is not exactly a typical personal computer.

This computer likely taught him a lot of technical things that he never would have learned without it in his possession. It was massive, but it was likely a giant Raspberry Pi for him and his friends, and it was the ultimate example of trash-into-treasure.

This phenomenon continues to this day. In fact, it’s even easier thanks to tools like Craigslist, eBay, Facebook Marketplace, OfferUp, and Shop Goodwill that are designed specifically for this.

So that’s why, to me at least, buying an old workstation makes sense.

Five reasons why you should consider buying old workstation gear

1. It can be had for very cheap, because its users upgrade often. Ever work in an office that has an IT department? You probably know that computers are swapped out on a three-year cycle, usually—though perhaps not if your office is rocking pre-butterfly MacBook Pros. This means that, after a few short years, those machines that they bought three years ago require an upgrade due to company rules about upgrading at a regular pace. Gotta keep up with tech trends in the industry, after all. Which means that machines that lived a good life are on the way to living a second life.
2. It’s often an inexpensive gateway to modern hardware. While there is some proprietary stuff that you’ll likely find in an old machine, you can easily add standard components like solid-state drives and even high-end graphic cards. There is even an enthusiast site out there, GreenPCGamers, that is dedicated to maxing out this old workstation and server hardware for purposes of … running lightweight productivity software, obviously. I’m not personally a heavy modern gamer, but I could certainly use my machine for that purpose at this point. And if I ever decide to upgrade, I can take what I’ve purchased in this machine and put it in a new one. (But why would I want to do that?!)
3. Because it uses old server gear, parts can be had for a steal. Some of the upgrades I did on my machine, including the processor and the RAM, are plentiful on the used market because they were originally used in places where their use cases are out of date. For a company running a server rack, eventually that older gear becomes less cost-effective as newer tech comes along that does the job more efficiently. This leads these old parts to flood the used market, particularly on sites like eBay. With the right amount of digging, old Xeon processors and ECC memory pulled from server racks can be had at rates way below their raw performance levels. Because it’s something the original audience doesn’t want, these things can be had for fairly cheap. (Case in point: Upon its 2013 release, the Xeon E5–2667 v2, one of the most robust chips my machine supports, had a list price of $2,057. Today, these chips sell in used form for as little as $180 on eBay—a more than 90 percent price decrease, a price $100 less than the roughly comparable AMD Ryzen 7 2700X, a popular chip with consumers these days.)
4. If you’ve never built a computer, it’s a great way to learn the ropes. Spend a while using laptops and suddenly want a home machine? An old workstation could be a great starting point, even if there’s no RGB lighting to be found. (You don’t need that!) The reason is simple: Because old OEM (original equipment manufacturer) gear was targeted at IT departments, everything is generally well documented and odds are good that, if you’re running into a problem, someone else has run into that wall themselves.
5. You’re recycling! Old workstations that aren’t made by Apple aren’t going to show up in a museum someday. They represent gear that is produced, sold, and forgotten about—in fact, much of the public isn’t even aware it exists, let alone that it can be extremely useful beyond its target market. It’s the very definition of what people think of when they talk about e-waste, and you’re doing your part to remove some of that from the market.

Of course, there are downsides here, let’s be honest

There are a lot of considerations when you do something like this. You have to know what you’re doing and you have to put your research in.

Among a few I can think of:

Things can be prone to break, considering these are old machines. (The supply isn’t unlimited.)

If you don’t know a lot about computers, you may find the onramp lane for something like this difficult to pull off, despite the fact that many machines formerly used in offices would be perfect for home users. A four-year-old i5 minitower that was mostly used for Excel by an office drone is often fine for someone who mostly surfs the web and does the same thing at home.

Sometimes listings and brand names can be quite obscure, which means that if you don’t know what you’re doing you can easily buy the wrong thing. Just to offer an example, Xeon processors have confusing naming conventions that are difficult to decipher without a chart. Likewise, there are a lot of different types of memory around, and one type doesn’t work with another. (In my case, “registered” and “unregistered” error-correcting memory can’t mix. Don’t cross the streams.)

The hardware itself can be proprietary, which might make it harder to upgrade components down the line. (Though this does have its benefits, depending on your machine: Only a few things on the inside of the HP Z420 actually require a screwdriver.)

And there are also considerations here from the perspective of power consumption. A big box that’s always plugged in will inevitably use more power than a tiny laptop, even if the big box can do a lot more. And older chips, reliant on prior technology, will of course be less efficient than something newer. Some older Xeon generations only reach modern levels of performance with multiple chips, for example, though you might get a better bang for your buck with something newer.

But if you can make the case for it, it might be worth your time. In my case, I was looking to have more of a desktop experience for times when I wanted slightly more horsepower than a laptop, and I also wanted a machine that could do virtualization when needed or desired. Finally, I wanted to have a desktop Hackintosh, but also wanted it to be good for, say, messing around with DOS or an old Amiga OS one day. (I would not be opposed to getting a floppy or Zip drive for this thing, just to keep up the retro vibes.)

I make no guarantees it will work for you, but it worked for me.

About once a week, I’ve taken to making a visit to my local Goodwill, which has a pretty robust electronics section. I often do trips like these to see if they inspire story ideas, or make me think about things I haven’t in a while.

What’s interesting, though, is that many of the things that land there are often not vintage computers, though you do see the occasional video game that’s more than 20 years old.

What shows up there are the Xbox 360s, the Playstation 2s, the Nintendo Wiis, and lots of old laptops that have decayed in value as their technology has turned from essential to old-hat. Many of these machines are broken—if your $400 Xbox 360 suffered from a red ring of death, it’s probably here. Some of their parts, such as old power bricks, flood the bins, selling for just a few dollars.

The weird and interesting things—such as the Windows tablet I bought a while ago—are more likely to disappear in a relatively orderly fashion.

But these utterly common machines? They face an uncertain fate. Some of them will find a home with someone willing to repair them. Maybe others will be eventually recycled and scrapped and turned into raw materials for new technologies that haven’t even been thought of yet. But these overly common devices, once the center of millions of lives, sit in decay.

I guess it’s part of the reason that I suggest that, if you’re looking for an upgrade, think about buying things that might be older, but could potentially have extra value for you. If it can be upgraded or user-serviced? Even better. We live in a world where not enough of our technology finds a place to live beyond its original context.

Electronic waste is a systemic problem. Giving a little extra life to tools that would otherwise be sitting in a landfill somewhere or might face issues going through the recycling process is something we should all do, especially when the use case allows you to avoid buying something totally new when you don’t have to.

It might even save you a bit of money in the process.