A homelab box that never sleeps runs 8,760 hours a year. So the spec that decides what it costs you is not the one on the box. It is the one nobody prints: how many watts it pulls sitting at the login prompt doing nothing.
I kept hitting this while shopping for a Proxmox node, so I put the measured numbers in one place. More on that at the end. First, why the spec sheet lies to you.
TDP is a thermal budget, not a power reading
TDP is the heat the cooler has to handle at full tilt. It is a design target for the heatsink, not a measurement of what the chip draws, and it says almost nothing about idle. Your homelab box spends 95%+ of its life idle, so the number that runs up the meter is idle wall power, and that number is never on the product page.
The arithmetic is unforgiving. One watt running continuously is 8.76 kWh a year. So the gap between a 7 W box and a 35 W box is not 28 watts, it is about 245 kWh a year, every year, for as long as the box is on. Plug in your own rate to get the dollars; the point is the gap compounds.
Where TDP actively misleads you
A few measured results from the dataset I'll link below, all from third-party wall-meter readings, not vendor claims:
The new N100 wave is genuinely low. A Minisforum UN100C measures 5 to 7 W at idle. Beelink, GMKtec and Trigkey N100 boxes land in the 6 to 10 W range. For a Pi-hole, a few containers and some light VMs, this tier is hard to beat on running cost.
AMD mini PCs idle far higher than their marketing suggests. A Minisforum UM790 Pro measures 25 to 45 W at idle. A Beelink SER6 Pro lands at 20 to 35 W. These are fast little machines, but if you picked one expecting "small box, small draw," the meter disagrees, and over a year that delta is real money.
Newer and higher-TDP is not lower-idle. A Dell OptiPlex 7060 Micro idles just over 18 W on its 65 W-TDP desktop chip. The older 7070 with a six-core part sits around 13 W, and the low-power "T" SKUs lower still. The CPU's TDP class predicted idle better than the model year did.
Core Ultra is not automatically frugal. An ASUS NUC 14 Pro with a Core Ultra part measures 15 to 28 W idle. A lower-end Core 3 100U variant of the same chassis idles under 10 W. Same box name, very different bill.
Used enterprise tiny desktops are the quiet sweet spot. Lenovo ThinkCentre M75q / M80q and HP EliteDesk 800 Mini units measure around 11 to 15 W idle, and they ship with ECC-capable platforms, real Intel NICs and proper firmware. An HP EliteDesk 800 G3 Mini idles around 12 W and tops out near 51 W on its 65 W adapter, so even loaded it stays modest.
Idle is the floor, not the bill
One more trap: idle is what the box draws doing nothing, not what your workload draws. A Minisforum MS-01 measures about 13 W bare-metal idle, but the same machine averaged around 52 W over a week running a full VM stack. Idle tells you the floor and lets you compare boxes fairly. Size your actual workload on top of it.
The other facts that aren't on the spec sheet either
Idle watts is the headline, but the same "vendor won't tell you" problem hits every deciding fact for an always-on box:
- ECC support. Most consumer mini PCs use non-ECC SODIMMs. If you care about ZFS or long-running VMs, this narrows the field fast, and it is rarely stated plainly.
- NIC chipset, not just speed. "2.5GbE" tells you nothing about the silicon. Intel i226 and i225 have different histories; some boxes ship different NIC silicon across batches, so the revision you receive matters. Check lspci -nn on arrival.
- The AMD reset bug. RDNA-class integrated GPUs (the UM790 Pro's iGPU among them) are subject to the AMD reset bug, which makes GPU passthrough flaky on Proxmox and sometimes unstable. If passthrough is your plan, this is a buy/no-buy fact.
- RAM ceiling. The DDR5 SODIMM boxes often cap at 96 GB regardless of what the slots imply.
What I built
I got tired of reconstructing this from a dozen forum threads per box, so I made IdleWatt, a small finder that ranks homelab mini-PCs and SFF boxes by measured idle watts, then layers ECC, NIC chipset, RAM ceiling, IOMMU/passthrough behavior and AMD-reset-bug status on top. Every deciding field carries a source citation and a date, or it is left out. It covers 30 boxes today, from N100 sticks to used enterprise tinys to the MS-01 class.
Finder: https://idlewatt.vercel.app
The underlying data is open. Every measured number traces back to a cited third-party source (wall meters, smartplug week-averages, Proxmox build write-ups), released CC-BY so you can check the work or reuse it:
Dataset: https://github.com/SolvoHQ/homelab-mini-pc-dataset
Disclosure: I built both. It is cited public information, not purchasing advice; SKUs, BIOS and NIC silicon change between batches, so verify the revision you actually receive before you rely on a passthrough or NIC claim.
If you have wall-meter idle readings for a box that is not in there yet, the dataset takes them. The whole point is measured numbers instead of TDP guesses.
Top comments (0)