Motor-start and surge ratings by model: the LRA table
Home battery surge ratings: six of the 17 models we track publish a motor-start rating in locked rotor amps. Nine publish a peak in watts, the rest none.
An induction motor at rest is close to a short circuit. Until the rotor turns it pulls several times its running current, and the size of that pull is not proportional to the watts the appliance uses once it settles. That moment is why peak power ratings exist on battery datasheets at all, and the unit that describes it on the appliance side is locked rotor amps. Most battery makers answer in a different unit.
This page is the reference table: what each of the 17 models on our comparison pages publishes for surge output and for motor starting, with the basis attached. Why the two columns behave so differently is the argument in our continuous versus peak power guide, which this page does not repeat. Every figure below is a manufacturer specification. We have not tested any of these products.
The table
| Model | Peak or surge output as published | Motor-start rating in amps |
| --- | --- | --- |
| Savant Power Storage 20 | On grid 30 kW for 300 ms and 28 kW for 25 s; off grid 24 kW for 400 ms, 19.1 kW for 1 s and 16 kW for 10 s | Not published |
| FranklinWH aPower 2 | Max Peak Output Power 15 kW at 10 s; Max Transient load capacity 25 kW at 1 s; Max Half-Wave Load Capability 5 kW | Max Load Start Capability 185 A LRA |
| Anker SOLIX X1 | Off grid Peak Output Power 9 kW for 10 s on the 10 kWh unit, 12 kW for 10 s on the 15 and 20 kWh units, 24 kW for 10 s on two 15 or 20 kWh systems | Row labelled LRA: 70 A/s for one system, 140 A/s for two |
| Fortress Power eVault Max 18.5 | Peak Discharge Rate 230 A, 12 kW for 30 minutes; Maximum Surge Power Rate 250 A, 12.8 kW for 5 s | Not published |
| SunPower SunVault | Max Backup Discharge Power 6.8 / 7.5 / 8.5 / 10 kW at continuous / 30 min / 10 s / 1 s on SunVault 13; 6.8 / 8.0 / 12.0 kW at continuous / 30 min / 10 s on SunVault 26 | Not published. Discontinued |
| LG Home 8 | Backup mode Max AC Power (Discharging) 9.0 kVA at 10 s, with Max AC Current 37.5 A at 10 s | Not published |
| Enphase IQ Battery 5P | Peak output current 32 A for three seconds and 25.6 A for ten seconds at 240 V, advertised as up to 7.68 kVA peak power | Power Start capability up to 48 A LRA |
| SolarEdge Home Battery 400V | Peak Output Power (for 10 seconds) 7,500 W | Not published |
| Tesla Powerwall 2 | Real Power peak 7 kW and Apparent Power peak 7.2 kVA, both at 10 s, off grid or backup only | Load Start Capability 88 to 106 A LRA |
| Enphase IQ Battery 10C | No peak kVA figure; peak output current 56 A for three seconds and 44.8 A for ten seconds at 240 V | Power Start capability up to 90 A LRA |
| Generac PWRcell | No peak power figure | Max AC Start Capacity 65 A at M3, 87 A at M4, 108 A at M5 and 130 A at M6; the PWRcell 2 Inverter is separately quoted at 130 LRA |
| Tesla Powerwall 3 | No peak power figure; a configurable off-grid maximum continuous discharge of 15.4 kW is published separately | Load Start Capability 185 LRA |
| BYD Battery-Box Premium HVS | Peak Output Current 50 A for 3 s at the battery terminals, against a Max Output Current of 25 A. A DC figure; the AC rating belongs to the paired inverter | Not published |
| EG4 PowerPro WallMount | No peak power figure; the BMS protection table publishes discharge trips above 205 A after 10 s and above 300 A after 3 s | Not published |
| Panasonic EverVolt 2.0 | No peak power figure | Not published |
| Qcells Q.HOME CORE | No peak or surge figure; battery Maximum power published at 5.5, 8.3 and 11.1 kW by size against a 7,608 VA inverter ceiling | Not published |
| sonnen sonnenCore+ | No peak power figure; publishes off-grid current thresholds instead, per stack: continuous 0 to 25 A, 5 s to disconnect at 30 to 36 A, 450 ms to disconnect at 36 to 50 A, immediate shutdown at 76 A peak | Not published |
Sources are the datasheets cited on each model page: Savant Power Storage 20 specification sheet, 009-2477-00, FranklinWH aPower 2 datasheet, Dec 2025, Anker SOLIX X1 AC-coupled datasheet, US, rev 2025-03, Fortress Power eVault Max 18.5 technical datasheet, SunPower SunVault ESS datasheet, RevC 2021, LG official spec page for the RA768K16A11, Enphase IQ Battery 5P datasheet, May 2025, SolarEdge Home Battery 400V datasheet, Jan 2026, Powerwall 2 datasheet, Enphase IQ Battery 10C datasheet, Feb 2026, Generac PWRcell 2 spec guide, Jul 2025, Powerwall 3 datasheet, BYD Battery-Box Premium HVS and HVM datasheet V1.7 EN, EG4 PowerPro WallMount spec sheet, ver. 1.0.1, Panasonic EverVolt Home Battery System datasheet, marked preliminary, rev 2023, Qcells Q.HOME CORE datasheet, H3S and H7S, May 2025 and sonnenCore+ technical specifications. All read August 19, 2026.
The basis notes that decide whether a number transfers
Several figures in that table carry conditions on the same page they appear on, and quoting the number without the condition is how comparison tables go wrong.
FranklinWH scopes its whole off-grid block to one configuration. The 15 kW peak, the 25 kW transient, the 5 kW half-wave figure and the 185 A LRA all carry the same footnote: "This value applies to a rated 11.5 kVA." The aPower 2 ships in four power ratings from 5 kW to 10 kW nominal, and the surge numbers belong to the top one.
Savant publishes different numbers on grid and off grid, and the headline is the on-grid one. The 30 kW at 300 ms that appears in every summary of this product applies while the utility is up, which is the condition under which you need it least. During an outage the ten-second figure is 16 kW.
Anker's peak figures come with a temperature and charge window. The datasheet footnotes them as valid at an ambient temperature of 50 to 131 degrees F with a state of charge greater than 35 percent, and separately notes that the 10 kWh system derates above 113 degrees F.
Generac states the household conditions its start capacity assumes. The figures are footnoted as "based on 2 kW of power consumption prior to A/C start with 12-100% SOC at an ambient temperature below 104 °F (40 °C)", with a second footnote that derating may occur above 104 degrees F and below 32 degrees F.
Tesla and Enphase both flag variability. Tesla footnotes the Powerwall 2 range with "Load start capability may vary." Enphase footnotes both of its LRA figures with "Power Start capability may vary."
LG's own spec page disagrees with itself. The backup mode table publishes 9.0 kVA at 10 seconds and a 7.5 kVA rated figure. The stacking table further down the same page reverses those labels for one unit. We record the backup mode table and flag the inconsistency rather than picking a side.
Three products publish a protection threshold instead of a surge rating
Three datasheets publish current limits that look like surge specifications and are not. They are the points at which the system protects itself, which is a different and in some ways more informative disclosure.
sonnen publishes an off-grid table with four rows per stack size: continuous operation at 0 to 25 A, disconnect within 5 seconds at 30 to 36 A, disconnect within 450 ms at 36 to 50 A, and immediate shutdown at 76 A peak, all for a single unit and scaling with the stack. Read as a surge envelope, that says a sonnenCore+ tolerates an overload for a stated time before it opens, which is closer to what you want to know than a bare watt figure. It is not labelled a peak power rating and we do not present it as one.
EG4 publishes a BMS protection table with discharge trips above 205 A after 10 seconds and above 300 A after 3 seconds. Those rows sit under a Discharge heading but are labelled "Over-Charging Current", which is the datasheet's own labelling and not ours. The PowerPro is a DC module in any case, so what your house sees is set by the paired inverter.
BYD publishes a Peak Output Current of 50 A for 3 seconds against a Max Output Current of 25 A, at a nominal 409.6 V on the HVS 10.2 tower. Those are battery-terminal DC figures, and we are not converting them into an AC kilowatt rating, because the inverter that decides what the house gets is not in this datasheet.
What ten of seventeen do not tell you
Ten of the models we track publish nothing in locked rotor amps and nothing in any other motor-start unit: Panasonic, LG, SolarEdge, SunPower, sonnen, Qcells, EG4, Savant, Fortress Power and BYD. An eleventh, Anker, prints a row labelled LRA whose value carries a per-second unit rather than plain amps, which is closer to useful than watts and is still not a figure you can set directly against a nameplate.
That absence is not evidence of a weak inverter. Savant publishes a 30 kW surge and no LRA. Fortress publishes 12.8 kW for 5 seconds and no LRA. Those are capable products whose makers chose a different unit. What the absence does mean is that answering the only question that matters here, whether your specific motor will start, requires a conversion and an assumption in the middle.
That conversion looks harmless and is not. Multiplying locked rotor amps by the supply voltage gives apparent power in volt-amps, but a motor at locked rotor has a poor power factor, so the real watts drawn are lower. Manufacturers do not publish a locked-rotor power factor. We use the current-derived figure where we have to, because inverter protection trips on current rather than on true power, which makes it the conservative side. A published LRA rating removes the guesswork entirely.
How to use the table against a nameplate
Three steps, in this order.
Find the largest motor you need to start, and read its nameplate. Compressor and pump nameplates print LRA directly. Franklin Electric publishes locked rotor amps for its submersible motors in its AIM Manual, running from 32.2 A on a 1/2 HP 230V two-wire motor to 66.2 A on a 1.5 HP, and the full comparison against real equipment is in our well pump guide.
Compare against a published LRA figure if the battery has one. No conversion, no assumption, same unit on both sides.
If the battery publishes only watts, use the duration nearest a motor start. That is the one to two second row, not the ten second row and certainly not a 30 minute row. The air conditioner version of this arithmetic, including what a soft start changes, is in our air conditioner guide.
And ask the off-grid question explicitly. Savant's off-grid surge figures are materially lower than its on-grid ones, and Tesla's Powerwall 2 peak is published as off grid or backup only. An outage is the condition you are buying for.
Every figure above is a manufacturer datasheet claim recorded with its published basis, or a recorded absence of one. Manufacturers revise datasheets and the same model number can carry different figures across revisions, so confirm the current document for the exact configuration you are quoted. Your own nameplate is the number that governs.