The best home battery backup for a grid blackout is a stack

The best home battery backup for a grid blackout is usually not the biggest battery you can afford. In a smart home, the first failure that matters is often smaller and dumber: the modem, router, hub, border router, switch, or Home Assistant host loses power, and every battery-powered sensor in the house becomes a device with nowhere useful to report.

That changes the buying order. Start by keeping the control layer alive. Then add enough portable capacity for the refrigerator and a few essentials. Only after that does a whole-home battery make sense as the center of the plan.

Size the blackout plan from watts, not product names

The useful question is not “which battery is best?” It is: what must stay alive, at what watt load, for how long? A typical smart-home network stack is not a huge load, but it is made of several small devices that fail at the worst possible time. HomeTechHacker’s smart-home UPS sizing guide gives practical ranges: modem 8–12 W, router 10–15 W, access point 8–12 W, smart-home hub 3–6 W, network switch 10–25 W, camera base or NVR 10–30 W, and NAS 20–60 W.[1]

Those ranges make the first design choice obvious. A modem, router, and hub may sit around 25–35 W. Add an access point and a switch and the same closet can become a 45–70 W load. Add an NVR and NAS, and it may no longer belong on the “keep alive at all costs” UPS unless surveillance and local storage are part of your outage plan.

Runtime math is just as important as load math. A practical estimate is:

usable runtime in hours = battery watt-hours × 0.85 ÷ device watts

The 0.85 factor is a rough allowance for conversion losses, not a guarantee. PiForz uses this method for router backup planning and gives the simple example of a 500 Wh power station running a 20 W router for about 21 hours. The same source also flags the trap in VA ratings: a 1500 VA UPS may hold only about 216 Wh, so VA should not be read as runtime.[2]

Run the numbers before reading any “best battery” list. A 35 W modem-router-hub stack on 216 Wh of actual battery capacity is not a day-long solution after losses. Using the same estimate, it is roughly 216 × 0.85 ÷ 35, or a little over five hours. If your outage pattern is one-hour utility flickers, that may be excellent. If your local storms regularly leave the block dark overnight, it is not enough.

Layer 1: keep the modem-router-hub stack from rebooting

The first layer should be boring. A small UPS or DC mini-UPS should carry the devices that make the rest of the smart home observable and controllable: modem or ONT, router, hub or coordinator, Thread border router if it is separate, and whichever switch is required to connect them. If your Wi-Fi access point is separate from the router, decide whether it belongs in the same survival group or whether one wired management path is enough.

This is the layer most whole-home backup discussions skip. Battery-powered Zigbee and Z-Wave sensors may still have power during an outage, but the useful path depends on the coordinator, hub, router, and automation host staying alive. The same idea shows up in practical compatibility planning: devices that work locally still need a powered local control path. That is why a wind or storm plan should separate “device has a battery” from “the system can still use the device’s state.” For more on that distinction, see which smart-home devices keep working during a wind advisory.

For Zigbee homes, this is not academic. The coordinator or hub is the mesh’s gateway to automations and apps. If it drops, the individual end devices may not be the weak point. The hub is. If you are still deciding where that responsibility should live, the hub and coordinator tradeoffs are covered in Zigbee hub requirements and device ecosystem planning.

A good Layer 1 setup has three traits. First, it has enough watt-hour capacity for the outages you actually get, not the optimistic runtime printed on a box. Second, it switches fast enough that the router and hub do not reboot during brief flickers. Third, it comes back cleanly when utility power returns.

That last part deserves more respect. A power plan that keeps the refrigerator cold but leaves Home Assistant wedged after a messy shutdown is still a failed smart-home plan. James Ridgway’s field report on home automation resilience notes that Raspberry Pi-style boards tend to restart automatically after power returns, while NUC-style Home Assistant hosts may need BIOS “power on after failure” settings or Wake-on-LAN configured to recover cleanly.[4]

The same recovery thinking applies to network order. If your modem takes several minutes to renegotiate service, your router boots faster, your hub boots faster still, and cloud-dependent integrations start timing out, the house may technically be powered but still not settled. A UPS on the whole network stack avoids many of those disorderly restarts because nothing in that control path ever saw the outage.

What belongs on the small UPS

• Definitely include: modem or ONT, router, smart-home hub or coordinator, and any border router or bridge required for local control.
• Usually include: one small switch if it connects the router, hub, access point, or Home Assistant host.
• Include only if needed: one access point, especially if the router is not also the Wi-Fi source.
• Question hard: NAS units, NVRs, camera bases, and large PoE switches. They may matter, but they can also consume the runtime that would have kept the control layer alive.

If the outlet is crowded with low-voltage wall warts, a DC mini-UPS can be cleaner than a conventional AC UPS, but only if voltage, connector size, and current ratings match the devices. If you cannot verify that, a conventional UPS is less elegant but usually less risky.

Layer 2: use a portable power station when the outage becomes a household problem

Once the network layer is stable, the next useful step is not necessarily a wall-mounted battery. It is often a portable power station in the 500 Wh to 2 kWh class. This is the layer that can carry the network stack longer, rotate through phone and laptop charging, run a few lights, and support refrigerator management without rewiring the house.

The same runtime formula applies, but portable stations make the difference between small continuous loads and large cycling loads more obvious. A router that draws 20 W is simple: a 500 Wh station is about a 21-hour estimate after typical losses.[2] A refrigerator is not that tidy because it cycles, and its draw depends on model, room temperature, door openings, and compressor behavior. The right way to size it is to measure energy use over a normal day or use the appliance’s documented consumption as a starting point, then leave margin.

Hands-on reports are useful here, as long as they are treated as reports rather than universal lab results. ZDNET describes a 286 Wh power station as enough to cover a full workday for the author’s permanently connected setup, and says roughly 1 kWh can keep modem and router loads running for days.[3] That does not prove your load will do the same. It does show why a small network load is worth protecting separately: it turns modest watt-hours into meaningful continuity.

For a smart-home owner, the portable station should not replace the small UPS. The UPS handles flickers and prevents reboots. The power station extends runtime and carries larger essentials. During a long outage, you can plug the UPS into the power station, or move selected loads over deliberately. What you do not want is the whole network stack dropping while you are hunting for cables in the dark.

A practical portable-station loading order

1. Keep the modem-router-hub UPS charged or powered through the station.
2. Run refrigerator intervals according to measured need, not continuously by habit.
3. Charge phones, a laptop, radios, and battery lights while larger loads are off.
4. Avoid adding high-draw comfort loads unless the station and inverter are sized for them.
5. Preserve reserve capacity for the night and for the reboot sequence after power returns.

Solar input can help a portable station, but it should be counted cautiously. Panel angle, shade, weather, smoke, winter sun, and charge-controller limits can all turn a vendor’s recharge chart into a best-case scenario. Treat solar as a way to stretch runtime, not as proof that a small station has become a whole-home system.

This is also where storm-prep habits matter. If you know a major weather event is coming, charging the station, testing the UPS, labeling the critical plugs, and trimming unnecessary loads beat improvising after the outage starts. The same kind of pre-outage routine appears in California smart-home winter prep for a strong El Niño, where UPS-backed networking is one part of a broader storm plan.

Layer 3: whole-home batteries are for circuits, days, and bigger consequences

A whole-home battery changes the problem from “what can I plug in?” to “which circuits should keep behaving as if the grid were still present?” That can be worth it if outages are frequent, long, unattended, or dangerous for the house. It is also where the numbers get large quickly.

As a sizing anchor, solar.com says a 10 kWh battery can power critical systems for at least 24 hours when air conditioning and electric heat are excluded.[5] EnergySage’s home battery guidance puts HVAC-inclusive backup on a different scale, with roughly 30–40 kWh needed when heating or cooling is part of the backup expectation.[6]

The Lawrence Berkeley National Laboratory simulation is encouraging but easy to overread. It modeled solar-plus-storage performance for three-day outages using 2022 simulations. In that setup, a 10 kWh system fully met basic backup needs in virtually all U.S. counties when HVAC was excluded. When critical load including HVAC was considered, the modeled system met about 86% of the load at 10 kWh and about 96% at 30 kWh.[7]

The caveat is the point. That result assumes solar recharging and separates out HVAC in the basic-load case. It should not be read as “a 10 kWh battery alone backs up a whole house for three days.” It says something narrower and more useful: if you define basic loads carefully and have solar available to recharge, 10 kWh can be a strong essential-circuits tool. HVAC changes the scale.

That is why the whole-home decision should begin at the critical-load panel, not at a battery brand page. Put the network stack, refrigerator, selected lighting, garage door opener if needed, sump pump if present, medical or accessibility equipment, and the minimum HVAC equipment you actually intend to support on the table. Then calculate. A battery that looks oversized for networking can still be undersized for electric heat, central air, a well pump, or multiple large appliances.

Do not use old incentive math in 2026

For a 2026 purchase, do not price a battery project as if the former federal 30% battery tax credit is still available. The credit expired for systems installed after December 31, 2025, so older net-cost examples can be stale in a way that materially changes the decision. State and utility programs, including programs such as California SGIP, are worth checking, but they should be treated as site-specific possibilities rather than assumed savings.

Battery or generator is a runtime question, not a loyalty test

Batteries and generators solve different outage problems. Battery systems are quiet, can switch automatically, and are well suited to electronics and essential circuits. Their hard limit is stored energy unless they can recharge. Generator setups can run as long as fuel is available, but they bring noise, exhaust, maintenance, fuel storage, and siting constraints. Backup-power guides from PowerOutage.us, Wirecutter, and EnergySage all frame the tradeoff around switchover, runtime, noise, cost, and operating constraints rather than one universal winner.[8][9][10]

For the network-first stack, that means a generator does not erase the need for a UPS. Even with a generator, the modem and router still need to ride through the gap before the generator starts, stabilizes, and carries the circuit. A UPS is the bridge. A battery station or whole-home battery is the quiet runtime. A generator, if you choose one, is the refueling strategy.

The stack that usually makes sense

For many smart homes, the most resilient setup is layered like this: a small UPS or mini-UPS dedicated to the modem-router-hub layer; a portable power station sized from measured loads for refrigerator support, charging, lights, and longer network runtime; and a whole-home or essential-circuits battery only when the outage history, unattended operation, medical needs, sump pump risk, HVAC requirement, or multi-day target justifies the installation.

The useful buying sequence is simple enough to write on painter’s tape and stick inside the network cabinet: measure watts, decide hours, convert watt-hours after losses, test switchover, test recovery. If the modem, router, hub, and host survive the first hour cleanly, the smart home remains a tool instead of becoming another thing to troubleshoot in the dark.

References

1. How To Choose The Right UPS For A Smart Home — HomeTechHacker
2. How Long Will a Battery Backup Run a WiFi Router? Complete Backup Guide — PiForz
3. 3 things I keep permanently connected to a power station — ZDNET
4. Surviving A Power Cut: Home Automation Resilience — James Ridgway
5. How Long Can a Battery Provide Power During an Outage? — solar.com
6. Best Home Batteries — EnergySage
7. Study Provides First-of-Its-Kind Assessment of Solar+Storage Backup Power Potential During Long Duration Power Interruptions — Lawrence Berkeley National Laboratory
8. Battery Guide — PowerOutage.us
9. Choosing the Best Home Backup Power Options — Wirecutter
10. Battery backup power vs. generators: which is right for you? — EnergySage