NAS Power Consumption: The Real Electricity Math
A typical home NAS costs a few dozen dollars a year in electricity — not nothing, but far less than most people fear. Rough classes: a 2-bay ARM or N100 box idles around 10–20W, a 4-bay x86 NAS with drives active runs about 30–45W, and each 3.5" hard drive adds roughly 4–6W idle / 7–9W active. The math: 30W running 24/7 ≈ 263kWh/year ≈ $32–42/year at typical US rates of $0.12–0.16/kWh. That's about one old incandescent bulb left on — and roughly 3–5× less than repurposing an old gaming PC as a server. All figures are estimates; your drive count, workload, and electricity rate move the number.
The one thing to know: NAS power draw is modest and worth estimating honestly. A typical 2-bay home NAS idles around 8–15W (drives spun down) and pulls roughly 20–35W with two active hard drives; a 4-bay box runs higher — very roughly 25–50W in normal use, more during a rebuild. Each 3.5" hard drive adds about 4–8W. As a rough estimate, a box averaging 30W runs continuously for about 260kWh per year — near a US average electricity rate that's on the order of a few dozen dollars a year (varies widely by your rate and drive count). Spin-down and scheduled power on/off cut this, but frequent spin-up/down adds wear; leaving drives running is often kinder to them.
How many watts does each class of NAS actually draw?
Manufacturers publish best-case idle figures; real-world draw depends on drive count and what's running. These are honest approximate ranges, at the wall, for common home configurations:
| Class | Idle (drives spun down) | Typical active | Rough annual cost* |
|---|---|---|---|
| 2-bay ARM / Intel N100 NAS | ~8–15W | ~10–20W (enclosure) + drives | ~$12–30/yr |
| 2-bay x86 NAS, 2 HDDs active | ~10–18W | ~20–35W | ~$25–45/yr |
| 4-bay x86 NAS, 4 HDDs active | ~15–25W | ~30–45W | ~$35–60/yr |
| Each additional 3.5" HDD | ~4–6W | ~7–9W | ~$5–11/yr per drive |
| Repurposed gaming PC as server | ~80–150W idle | 100W+ typical | ~$95–190+/yr |
*Rough estimates assuming 24/7 operation at US $0.12–0.16/kWh. These are not measurements of specific models — spec-sheet idle numbers, drive models, add-in cards, and your utility rate all shift them.
What's the formula for annual electricity cost?
The whole calculation is one line, so you can localize it to your own rate:
Average watts × 8,760 hours/year ÷ 1,000 = kWh/year, then × your $/kWh rate.
Worked example: 30W × 8,760h = 262,800Wh ≈ 263kWh/year. At $0.12/kWh that's ~$32/year;
at $0.16/kWh, ~$42/year. (Shorthand: watts × 8.76 = kWh/year.)
For context, ~263kWh/year is in the neighborhood of a single old 60W incandescent bulb left on half the day, or a fraction of what a refrigerator uses. If you're in a high-rate region (California, parts of Europe at $0.30–0.40+/kWh), the same 30W box costs roughly $80–115/year — the formula is the point, not our number.
What actually moves a NAS's power draw?
Four things dominate, in roughly this order:
1. Drive count. At ~4–9W each, four spinning 3.5" drives can out-draw the enclosure's board itself. This is a quiet argument for fewer, larger drives when total capacity is equal — and it's why our bay-count guide tells you not to fill bays you don't need yet.
2. CPU class. A current Intel N100 (as in the UGREEN NASync line) or N95 (TerraMaster F4-424) idles meaningfully lower and does more per watt than the older Celeron J-series in entry Synology boxes. All of the x86 boxes we list are low-power parts, though — the class gap between any of them and a desktop CPU is far bigger than the gap between them.
3. Spindown policy. Covered below — it's the biggest number you directly control, and the most argued-about.
4. Add-ins and load. A 10GbE card can add several watts continuously, NVMe cache drives add a little, and RAID scrubs, rebuilds, and transcoding push the CPU and every drive to their active numbers for hours at a time.
Should you spin down your drives? (the eternal forum debate)
HDD hibernation is the most contested topic in home-NAS power, and both camps have a real point — so here are both, honestly.
The case for spindown: a hibernating drive drops from ~5–8W to under 1W. On a 4-drive box that idles most of the day, aggressive spindown can cut 20W+ from the average — real money at high electricity rates, and a genuinely large fraction of a mostly-idle NAS's total draw.
The case against: every spin-up is the highest-stress moment in a drive's day (motor surge, head load), and drive makers rate load/unload cycles in finite numbers. Frequent spin-up/down cycling adds wear that continuous spinning avoids — NAS-rated drives are explicitly built for 24/7 operation. There's also a usability tax: 5–15 seconds of dead air every time you open a folder against sleeping drives. And in practice, Docker containers, media indexing, and health polling often wake drives so constantly that hibernation never actually engages — you get the wear without the savings.
The honest synthesis: match the policy to access patterns. A daily-driver NAS running services: leave the drives spinning. A cold-archive or backup-target box touched weekly: spindown — or better, a power-on schedule — saves real money with few wake cycles. There is no setting that wins both ways.
NAS vs repurposed gaming PC: the honest DIY power tradeoff
"Just use your old PC as a server" is common DIY advice, and it hides a recurring cost. An older gaming tower or desktop typically idles at 80–150W — the desktop CPU, discrete GPU, and oversized PSU all burn power doing nothing. Against a 30W NAS, that's 3–5× the electricity: very roughly an extra $50–150 per year at typical US rates, every year, forever (estimates). Over a few years the "free" server can quietly cost more than a purpose-built low-power box — which is exactly why efficient mini-PC and N100 builds dominate modern DIY homelab advice. The full tradeoff (hardware value, ZFS, control, and this power gap) is in our DIY TrueNAS vs Synology explainer.
The other comparison people mean when they ask this question is cloud storage: a NAS's electricity is part of its true ownership cost, and it belongs in the math against a subscription — see NAS vs cloud storage cost.
What size UPS does a small NAS need?
Small NAS, small UPS. The one-line VA math: a 40W NAS is roughly 45–50VA of load, so even a modest 600–900VA consumer UPS is loaded under ~10% and will run it for tens of minutes (rough estimate) — far more than the minutes needed for a clean shutdown. Don't buy runtime; buy the USB data connection: plug the UPS's USB cable into the NAS so it can shut itself down gracefully when the battery kicks in. Protecting the array from an unclean power-loss mid-write is the entire job.
How do you measure your NAS's real draw?
Don't trust spec sheets — measure. A plug-through watt meter (Kill A Watt-class usage monitor) or an energy-monitoring smart plug between the NAS and the wall shows live watts and, more usefully, accumulated kWh. Log a full week — a spot reading during a scrub or a quiet idle hour will mislead you in either direction — then multiply the weekly kWh by 52 and by your utility rate from an actual bill. That number, not ours, is your real annual cost.
Which of our picks are the most power-efficient?
Among the boxes on our main list, the modern low-power CPUs win: the UGREEN NASync DXP2800 and DXP4800 (Intel N100) and the TerraMaster F4-424 (Intel N95) use the newest, most efficient silicon here. The Synology DS224+ is frugal too — its older Celeron J4125 is slower per watt but still a low-power part in a 2-bay chassis with fewer drives to feed. The DS923+ and the 4-bay QNAP boxes draw more mainly because four drives draw more than two; the enclosure differences are smaller than the drive-count difference. Practical rule: bay count you actually populate, then CPU generation, decide your bill — not the brand.
Frequently Asked Questions
As a rough estimate, a typical home NAS averaging 30W draws about 263kWh per year (30W × 8,760 hours ÷ 1,000). At typical US residential rates of $0.12–0.16/kWh that's roughly $32–42 per year. A small 2-bay ARM or N100 box averaging 15W is about half that; a repurposed gaming PC idling at 100W is 3–4× more. Multiply your box's average watts by 8.76 to get kWh/year, then by your local rate.
Rough classes: a 2-bay ARM or Intel N100 NAS idles around 10–20W; a 4-bay x86 box with drives active runs about 30–45W; each 3.5" hard drive adds roughly 4–6W at idle and 7–9W when active. Rebuilds, scrubs, and heavy transcoding push draw higher temporarily. These are estimates that vary by model, drive count, and workload — a plug-in watt meter gives you your real number.
It's a genuine tradeoff, not a free win. Spindown can cut a mostly-idle NAS's draw by 4–6W per drive, but every spin-up adds mechanical wear and a 5–15 second delay before files respond — and background services (Docker, indexing, health checks) often wake drives so frequently that hibernation never sticks. For a NAS accessed daily, many experienced users leave drives spinning; for a cold-archive box touched weekly, spindown or scheduled power-off saves real money.
Usually, by a wide margin. A purpose-built NAS averages roughly 20–45W, while an older gaming PC or tower server commonly idles at 80–150W — call it 3–5× the electricity, or on the order of $80–150+ more per year at typical US rates (rough estimates). That gap is a core part of the DIY-vs-turnkey decision: DIY hardware is cheaper up front per unit of performance, but an inefficient DIY box can give some of that back on the power bill every year.
No — a small one is fine. A NAS drawing 30–45W is a tiny load; a modest 600–900VA consumer UPS typically runs it for well over 15–30 minutes, which is plenty for a clean automatic shutdown (rough estimate: runtime ≈ enough when your load is under ~10% of the UPS's rating). What matters more than size is USB connectivity to the NAS so it can shut itself down safely before the battery dies — an unclean power loss mid-write is exactly what a UPS exists to prevent.
Use a plug-through watt meter (Kill A Watt-class, or a smart plug with energy monitoring) between the NAS and the wall. Log it across a full day, because a NAS's draw swings between idle, active, and scrub/rebuild states — a single spot reading undersells or oversells it. The meter's kWh-accumulated mode over a week, multiplied out to a year and by your utility rate, gives a far more honest annual cost than any spec-sheet number.
Keep decoding
All wattage and cost figures on this page are rough estimates from manufacturer specifications and aggregated owner measurements, not lab tests of specific units, and electricity rates vary widely — use the formula with your own rate and a watt meter for your real number. Electricity estimates are not Amazon product pricing. How we evaluate →