Tested UPS Runtimes for Smart Home Backup Power

For smart home backup power during outages, the number that matters is not the UPS’s biggest printed VA rating. It is how long that UPS can run your actual network stack: modem, router, smart home hub, switch, and whatever else you decide is worth keeping alive.

That distinction shows up sharply in Wirecutter’s 2026 UPS testing. At a defined 20W modem-plus-router load, the CyberPower LE1000DG ran for 3 hours, the CyberPower CP1350AVRLCD3 ran for 4.5 hours, and the Amazon Basics Standby 800VA ran for 1.5 hours. Put those same units under a 300W load, and their runtimes dropped to 12 minutes, 21 minutes, and 6 minutes, respectively.[1]

That is the pre-purchase lesson. The same battery box can look roomy or useless depending on what you plug into it. A modem and router are a very different load from a desktop PC, monitor, speakers, and external drives. If the goal is to keep automations, camera notifications, remote access, and hub logic alive during an ordinary outage, the UPS should be judged at network-stack wattage, not at some imagined household load.

Start with watts, not VA

VA ratings are still useful for understanding the class of UPS you are shopping in, but they are a poor shortcut for smart-home runtime. A 1000VA-class unit does not mean “three hours” or “four hours” by itself. It means nothing practical until the load is attached.

Wirecutter’s tested contrast is useful because the load definition is explicit. At 20W, the 1000VA-class CyberPower LE1000DG reached 3 hours; at 300W, it reached 12 minutes. The larger CyberPower CP1350AVRLCD3 reached 4.5 hours at 20W and 21 minutes at 300W. The smaller Amazon Basics Standby 800VA reached 1.5 hours at 20W and 6 minutes at 300W.[1]

Those are not universal runtimes for every UPS with a similar number on the box. They are measured results for named models, under named loads, in Wirecutter’s 2026 testing. That boundary matters. A different UPS chemistry, battery age, inverter design, or load mix can change the result. But the pattern is exactly what a smart-home buyer needs to see before spending money: runtime follows watts.

Estimate your smart-home stack before choosing the UPS

A typical smart-home core is small compared with the devices people usually imagine when they hear “backup power.” The part that keeps the house reachable is often a modem, router or mesh base, one or more hubs, a small Ethernet switch, and maybe a NAS or NVR. The NAS or NVR is usually where the tidy little power budget starts to stretch.

HomeTechHacker’s UPS sizing guide gives practical estimated ranges: modem 8–12W, router 10–15W, smart home hub 3–6W, switch 10–25W, and NAS 20–60W.[2] Treat those as starting estimates, not lab measurements of your gear. A cable gateway, a Wi-Fi 7 mesh base, a PoE switch, and a multi-bay NAS will not behave like the smallest devices in those ranges.

For a first-pass estimate, add only the devices that must stay on during the outage. A basic stack using HomeTechHacker’s ranges might look like this:

Using the low-to-mid parts of those ranges, a modem, router, hub, and small switch often lands around the 25–60W zone. Add a NAS or NVR and the stack can move closer to 50–120W, depending on drives, processor load, and whether the switch is powering other devices. A practical planning range of about 25–100W fits many smart-home network cores while leaving room for a hub and switch, but it should still be checked against the devices in front of you.

If you can measure, measure. A plug-in power meter between the wall and a power strip gives a better answer than any estimate. Let the network run normally, note the watts, then repeat while cameras are active, the NAS is awake, or the hub is busy after a reboot. The UPS does not care that a device is “small.” It cares about the total draw at the battery outlets.

Pick the outage window you are actually buying for

The right target is not infinite runtime. It is the outage window where keeping the network alive changes the outcome: alerts still go out, cameras stay reachable, automations do not lose their controller, and you do not spend the evening restarting half the house.

For U.S. context, the Energy Information Administration reported that customers averaged about 2 hours per year of electric interruptions excluding major events over 2013–2022, and 5.5 hours in 2022 when all events were included.[3] In 2024, EIA reported that U.S. electricity customers experienced 11 hours of interruptions on average, with major events accounting for about 80% of outage hours; state averages varied widely, including about 53 hours in South Carolina.[4]

That national data cuts both ways. It supports buying enough runtime for ordinary interruptions rather than pretending every home needs a whole-home battery. It also warns against using the national average as a promise. If your local outage history includes hurricanes, ice storms, wildfire shutoffs, or repeated utility failures, a two- to four-hour UPS plan is a continuity layer, not a resilience plan.

For a normal network stack, the tested Wirecutter numbers make a 1000VA-class UPS a plausible short-outage choice. A CyberPower LE1000DG ran 3 hours at the 20W modem-plus-router load, while the CP1350AVRLCD3 ran 4.5 hours at that same load.[1] If your measured stack is 50W or 80W instead of 20W, do not borrow those runtimes directly. Use them as evidence of the direction: small loads can run for hours; larger loads consume that margin quickly.

Decide what earns a battery outlet first

A UPS usually has more outlets than you should actually fill during an outage. The backed-up outlets should go to devices that keep the network and control plane alive. The surge-only outlets can protect devices that do not deserve battery runtime.

Manufacturer guidance from Anker SOLIX puts essential communications and networking equipment high on the priority list for home UPS use.[5] That is vendor guidance, not independent runtime testing, but the priority order is sensible for a smart home: keep the path from utility service to local control intact before adding convenience loads.

1. Power the modem, fiber ONT, or gateway first. Without it, remote access and cloud-dependent alerts may fail even if the router is alive.
2. Power the router or primary mesh node next. Local Wi-Fi, phone access, and many hub integrations depend on it.
3. Power the smart home hub or controller. That includes Home Assistant hardware, Hubitat, SmartThings hubs, alarm bridges, or Thread border routers if your setup depends on them.
4. Power the switch only if it carries critical wired links. A switch feeding the hub, ONT, access point, or camera bridge is different from a switch feeding a TV cabinet.
5. Add NAS or NVR power only after checking the runtime tradeoff. Storage can be important, but it is often the load that turns a comfortable UPS runtime into a short one.

This is also where Home Assistant users should think about the hub itself. A tiny dedicated box, an old laptop, and a more power-hungry mini PC can have very different outage behavior. If you are still choosing hardware, our guides to Pine64 Home Assistant hub alternatives and using an 8GB laptop as a smart home hub are worth reading before you assume every controller has the same power budget.

Match your load to tested runtime, then leave margin

Once you have a wattage estimate and target outage window, the buying process gets much cleaner. You are no longer asking whether a UPS is “big enough” in the abstract. You are asking whether a named unit has credible runtime data near your load.

A reasonable pre-purchase worksheet looks like this:

• Measured or estimated load: modem + router + hub + switch = your baseline watts.
• Optional load: NAS, NVR, extra mesh node, PoE switch, or camera bridge.
• Target runtime: for example, 1 hour, 2 hours, 4 hours, or enough time for a graceful shutdown.
• Runtime evidence: third-party test, manufacturer runtime chart, or both, with the load clearly stated.
• Margin: extra capacity for battery aging, higher real-world draw, and the temptation to plug in one more device later.

If your modem, router, hub, and switch measure 40W, a 20W test result is encouraging but optimistic. If a model barely meets your runtime target at a lighter tested load, it is not the right fit unless you are comfortable with less runtime than the test suggests. If your stack is 80W with a NAS awake, compare against runtime curves or tests closer to that load rather than leaning on the 20W modem-router number.

Battery age deserves some pessimism. UPS batteries lose capacity over time, and many small UPS units are bought once and ignored until the first outage proves they are tired. Buying with margin is less glamorous than buying the largest VA number, but it is the part that keeps the network from dropping at the ninety-minute mark.

The few UPS features that matter for hubs and routers

Runtime is the main decision, but a few features affect whether the outage is invisible to your gear or turns into a reboot storm.

Transfer time

A standby or line-interactive UPS needs a short moment to switch from wall power to battery. Wirecutter notes that its UPS picks have 4–8 ms switchover ratings, short enough that routers and hubs should not reboot during the transfer.[1] This is one of the reasons a proper UPS is different from manually moving plugs to a battery after the outage has already started.

AVR for brownouts

Automatic voltage regulation, or AVR, helps when voltage sags or rises without fully cutting out. Wirecutter highlights AVR as useful in brownout-prone areas.[1] For smart-home gear, that can matter as much as full blackout runtime, because flaky voltage can cause modems, hubs, and switches to restart even when the lights only flicker.

Modified sine wave is usually fine for this load

Many affordable UPS units output a modified sine wave on battery. Wirecutter’s guidance treats that as fine for networking gear, while noting that sensitive audio equipment and medical devices are different cases.[1] For a modem, router, hub, and small switch, pure sine wave output is usually not the feature to overpay for before you have solved runtime.

Monitoring is useful after the UPS is installed

If you already run Home Assistant, UPS monitoring can turn backup power from a black box into a visible part of the house. Home Assistant’s Network UPS Tools integration can expose UPS status and sensors through NUT, depending on the UPS and NUT setup.[6]

That does not mean you should buy a larger UPS just to get more dashboard data. It means that after the right-size UPS is installed, Home Assistant can help you see battery state, detect when the house is on UPS power, and trigger automations such as shutting down a NAS before the battery is exhausted. For a procedural pre-storm drill, pair this buying guide with Smart Home Storm Prep With Surge Protection and Backup Power.

Where a UPS stops being the answer

A UPS is excellent at one job: carrying a small, important load through a short outage or power disturbance without interruption. It is not a whole-home battery, and it is not a guarantee that the internet outside your home still exists.

If the ISP’s upstream cabinet, neighborhood node, or fiber equipment loses power and has no working backup, your modem and router may stay alive while the internet connection disappears. Local automations may continue, local camera recording may continue if the recorder has power, and local control may still work. Cloud services and remote access may not.

The UPS boundary is also crossed when the goal changes from network continuity to comfort or multi-day survival. Refrigerators, medical equipment, sump pumps, HVAC, induction cooking, and whole-home circuits belong in a different sizing conversation. Portable power stations and whole-home batteries are the next tier once the UPS can no longer cover the load or the outage window.

For the smart-home core, the buying decision stays refreshingly specific: measure or estimate the stack, choose the outage window, compare against tested runtimes at similar loads, and avoid treating VA as the answer. A 1000VA-class UPS can be enough for modem-router-hub continuity during many short outages, but only if the watts and runtime line up with the house you actually have.

References

1. Best Uninterruptible Power Supply UPS, Wirecutter
2. How To Choose The Right UPS For A Smart Home, HomeTechHacker
3. U.S. electricity customers averaged five and one-half hours of power interruptions in 2022, U.S. Energy Information Administration
4. U.S. electricity customers experienced 11 hours of power interruptions in 2024, U.S. Energy Information Administration
5. Essential UPS Home Applications, Anker SOLIX
6. Network UPS Tools (NUT), Home Assistant