The topic Your home lab’s PSU is a ticking time bomb, and it should be the first thing you… is currently the subject of lively discussion — readers and analysts are keeping a close eye on developments.
This is taking place in a dynamic environment: companies’ decisions and competitors’ reactions can quickly change the picture.
A power supply is one of the easiest things to forget about, especially in the process of repurposing an old PC. If the thing turns on, that’s good enough for some tinkering, right? Then tinkering turns into running a home lab with a few light services, which turns into some load-bearing ones like a DNS sinkhole. In a matter of months, you’ve turned that old gaming PC into your NAS, media server, and AI inference box that you rely on daily.
Underneath it all, that old PSU is still acting as the foundation. Just because it’s still reliably supplying power at this moment, doesn’t mean it will continue to do so. A power supply failure is the one event that can take the whole machine offline at once and, in the worst case, damage the components it was feeding on the way out.
A PSU doesn’t always fail politely. The protection circuitry inside a good unit exists purely so that any fault becomes a clean shutdown of the system and nothing more. When that protection is solid, the usual outcome of a dying PSU is simply a PSU that no longer turns on, but there’s a good chance that the old PC you repurposed doesn’t have top-of-the-line protection.
In the worst-case scenario, a dying PSU that loses regulation can push a voltage that’s out-of-spec downstream to your other valuable components, and this could outright kill your motherboard, CPU, drives, and anything else that’s connected. It’s not the most common outcome, but the risk is absolutely real. Even setting damage aside, a dead PSU means the entire box goes dark. On a storage box, a power event can cause a whole host of issues, and on a box you’re hosting services on, everything depending on it goes offline at the same moment and stays down until a replacement is sourced.
Capacitor aging is the most common way power supplies die, and it’s governed by heat and time. The electrolytic capacitors inside a PSU degrade faster the hotter they run, and the well-worn rule of thumb is that capacitor life roughly halves for every 10°C rise in operating temperature. In its first life inside a gaming PC, this supply ran a few hours a day and sat cold overnight. In a home lab, it runs continuously, which is a meaningful difference. What wears a capacitor is hours spent hot under load, not necessarily the years it sat on a shelf.
A single year of around-the-clock operation is roughly 8,760 powered hours; a decade of a couple of hours of gaming a day lands in a similar range. A year or two of home lab duty can rack up as much real wear as the entire gaming life that came before it, and it does so continuously, with less time to cool. Budget and bundled gaming PSUs make this worse, and unless you were the original buyer and prioritized the power supply, there’s a high likelihood that the PSU within is not truly designed for always-on use and shipped with lower-grade capacitors.
A computer power supply converts wall power to DC at less than 100 percent efficiency, and the rest becomes heat. Older units often sit at lower efficiency tiers than current ones, and they’re the least efficient at the very low loads where a home lab spends most of its time. If you used a really high quality PSU in your old PC, and you’re repurposing it for your home lab now, this might not apply to you as much.
A PSU tends to be most efficient at around half its rated load and noticeably less so down near idle loads, so a machine drawing a few dozen watts most of the day is operating in the worst part of its curve. The dollar figure here is modest and not worth overstating, but it compounds quietly across thousands of hours of runtime.
If you think you should probably replace it, it’s probably well past due. The instinct when replacing a PSU is to buy big, and for a home lab that instinct is usually wrong. These machines draw far less than the gaming rigs they came from, often a few dozen watts at idle where they spend most of their lives, so a modestly sized unit is the better call. It runs closer to the efficient middle of its curve, stays cooler, and costs less up front than the 850-watt supply nobody needs. There’s one asterisk to that, which I’ll get to.
From there, a reputable 80 Plus-rated unit is the baseline, with one caveat: the badge describes efficiency, not build quality, and a credible third-party review tells you more about whether the protection actually works than the sticker does. I purchased a 750 watt unit that’s 80-Plus Gold certified, which is definitely on the high-end in terms of max wattage, but I plan to run a GPU for AI inference, and the extra headroom is necessary here. That’s the asterisk: if you’re running one or multiple GPUs in your home lab or plan to in future, aiming high is absolutely necessary.
If you repurposed an old PC for your home lab, it’s worth taking a closer look at the component that powers everything else. The point isn’t to panic and rip the supply out tonight, but if the line has shifted from “it turns on” to “it’s on 24/7 and powers core services”, it’s probably time to start looking for a replacement. Buying big is useless unless you plan on running power-hungry components or a very large array of storage, and you’d be perfectly happy with a lower-wattage unit from a reputable brand.