Why Your Mini PC's "6W" Spec Sheet Number Is Never What It Actually Draws
Every N100 mini PC spec sheet advertises a 6W TDP, and it’s technically accurate for the chip alone, measured at the wall outlet, with RAM, SSD, network controller and power supply losses included, that same system draws roughly 7-12W idle. For a machine running 24/7, that gap between the marketed number and the real one is the actual figure your electricity bill cares about, and it’s genuinely measurable with a £15-£19 (US$20-US$25) tool rather than something you have to guess at.
Key takeaways
- TDP describes the CPU chip alone, not the system around it RAM, SSD, network controller and power-brick inefficiency at low load all add real watts a spec sheet’s single TDP number never accounts for.
- The real gap is a few watts, not a small rounding error Independent wall measurements put a ‘6W’ N100 system at 6-12W idle, and higher-tier chips often idle well above their own rated TDP too.
- At 2026 US electricity rates, roughly £1.19 (US$1.60) per watt of continuous draw per year is the quick math A 10W idle difference between two otherwise similar mini PCs compounds to about £12 (US$16)/year, real money for a homelab running several boxes, genuinely trivial for one machine you use a few hours a day.
What TDP Actually Measures (and What It Deliberately Doesn't)
TDP (Thermal Design Power) is a chip manufacturer's rating for how much heat the CPU package itself is designed to dissipate, it's a cooling-design specification, not a promise about what the complete system draws from the wall. Everything else in the box adds its own draw on top: DDR4/DDR5 RAM typically adds 1-3W depending on capacity and DIMM count, an NVMe SSD adds a small but real amount, a 2.5GbE network controller adds more than people expect, and, often the most overlooked factor, the external power brick itself is rarely more than 80-88% efficient at the low power levels a mini PC actually draws at idle, meaning a real slice of your wall reading is simply the brick converting AC power to DC and wasting some of it as heat in the process.
Independent reviewers who actually measure this at the wall rather than quoting the spec sheet consistently find the same pattern: a '6W TDP' N100 system draws roughly 6-12W idle in practice, and interestingly, higher-tier chips often show a proportionally larger gap, some higher-wattage rated chips have been measured idling above their own rated TDP once the full system and BIOS power states are accounted for. The N-series specifically tends to be the most honest about its rating in practice; more powerful chips generally show a wider gap between marketed and measured numbers.
A basic plug-in power meter (a Kill-A-Watt-style device or an energy-monitoring smart plug) measures your specific unit’s real draw in minutes. For a machine running 24/7, this is worth doing once rather than trusting either the manufacturer’s TDP or a stranger’s review of a different unit.
The Real Numbers, by Chip Class
|
Most efficient, closest to rated TDP
Intel N100/N150 class
|
Moderate idle, strong performance-per-watt
AMD Ryzen U-series (e.g. 7-8000U)
|
Higher idle, highest performance ceiling
AMD Ryzen H/HS-series (e.g. Ryzen 9 8945HS)
|
|
|---|---|---|---|
| Typical measured idle (system, at the wall) | 6-12W | 7-10W | 15-22W |
| Typical measured full load | 20-30W | 25-35W | 77-79W (sustained stress test) |
| Approx. annual electricity cost, 24/7 idle, 2026 US average rate | ~£10-£14 (US$13-US$19) | ~£8-£12 (US$11-US$16) | ~£18-£26 (US$24-US$35) |
| Gap vs. spec-sheet TDP number | Smallest, closest to rated TDP | Moderate | Often the widest gap |
| Check Price | Check Price | Check Price |
Annual cost figures use the rough shortcut of £1.19 (US$1.60) per continuous watt per year, based on 2026 US average residential electricity rates around 18.34 cents/kWh, your actual rate varies by region and utility, so treat these as directional rather than exact for your specific bill.
When This Actually Matters (and When It Genuinely Doesn't)
Deciding how much to actually care about idle power
Idle power only compounds into real money on an always-on machine, a server, NAS, or home automation box.
A few watts per box is trivial once; it compounds meaningfully across a multi-box homelab.
The dollar impact of any given wattage difference varies significantly by region.
Manufacturing variance and BIOS/firmware differences mean your exact unit may differ from a reviewer’s sample.
Who Should Weight This Most Heavily
- A wall meter costs £15-£19 (US$20-US$25) and gives you an exact, unambiguous answer for your unit
- N-series chips genuinely deliver the lowest 24/7 running cost among mini PC options
- The annual cost difference, while real, is modest in absolute terms for most single-machine use cases
- Spec sheet TDP numbers are technically accurate but practically misleading for system-level planning
- Higher-performance chips show the widest gap between rated and real-world idle draw
- Power brick efficiency, an easily overlooked factor, meaningfully affects the final number
Exploring the wider hardware category
See our full hardware guide for mini PCs, storage, networking and accessories across every use case.
Our Sources
Where this comes from
The wall-measured wattage figures and cost math here are drawn from independent reviewer wall-outlet measurements (including data referencing TechPowerUp, NotebookCheck and ServeTheHome testing) and multiple independent 2026 homelab power-measurement guides, cross-checked for consistency given the genuine unit-to-unit variance in this kind of measurement.
-
Independent wall-measurement data reviewed
Multiple independent 2026 sources citing real reviewer wattage tests, not manufacturer TDP specifications alone.
-
Cost math cross-checked against 2026 US average electricity rates
The ~£1.19 (US$1.60)/watt/year shortcut is based on the 2026 EIA average residential rate, clearly noted as regionally variable.
-
No claims of our own wall-meter testing
We have not personally measured these specific units; figures are drawn from and attributed to independent reviewer testing.
Frequently Asked Questions
Frequently asked questions
Why does a '6W' mini PC actually draw more than 6 watts?
The 6W TDP figure describes the CPU chip’s thermal design rating alone. The complete system, RAM, SSD, network controller, and an external power brick that’s rarely more than 80-88% efficient at low load, adds real watts on top, typically bringing real wall-measured idle draw to 6-12W for that chip class.
How much does mini PC idle power actually cost per year?
Using a rough 2026 US average rate shortcut of about £1.19 (US$1.60) per continuous watt per year, a 10W idle mini PC costs roughly £12 (US$16)/year to run 24/7, while a 20W system costs around £24 (US$32)/year, your actual cost depends on your specific local electricity rate.
How do I measure my mini PC's real power draw?
A basic plug-in power meter (a Kill-A-Watt-style device or an energy-monitoring smart plug), costing roughly £15-£19 (US$20-US$25,) plugged in between the wall outlet and your mini PC’s power brick, gives you an exact reading for your specific unit within minutes.
Which mini PC chip class has the lowest real-world idle power?
Intel’s N-series (N100, N150) consistently measures closest to its rated TDP and shows the lowest real-world idle draw among common mini PC chip classes, typically 6-12W at the wall for a complete system.
Does idle power consumption matter if I only use my mini PC a few hours a day?
Generally not much, the dollar impact of a few watts’ difference is genuinely trivial if the machine is off or asleep most of the time. It matters specifically for 24/7 use cases like a home server, NAS, or always-on homelab box.
Final take
- TDP describes the CPU chip alone, not the complete system's real draw
- N-series chips measure closest to their rated TDP; higher-performance chips show wider gaps
- At 2026 US rates, roughly £1.19 (US$1.60) per continuous watt per year is the quick cost shortcut
A mini PC’s advertised TDP is an accurate chip-level cooling specification and a genuinely misleading number for planning real-world 24/7 running costs, the actual wall draw is consistently a few watts higher once RAM, storage, networking and power-brick losses are included, and that gap widens further on higher-performance chips. For a single occasionally-used machine, the difference is trivial. For a 24/7 server or a multi-box homelab, it’s a real, easily measurable, and worth-checking number, and a £15-£19 (US$20-US$25) wall meter answers it definitively for your specific unit in minutes.
