Solar panels don't stop working when the sun ducks behind a cloud. They slow down. That's a different problem with different solutions, and one most homeowners learn the hard way when a week of gray weather lands in November and the monitoring app shows production at a fraction of what it managed in July. Here's what actually happens, and how a sensibly designed system handles it.
TL;DR: Heavy overcast cuts solar output to roughly 10-25% of peak, not zero. NREL's National Solar Radiation Database tracks diffuse irradiance around 100-200 W/m2 under thick cloud versus the 1,000 W/m2 peak reference. A grid-tied system without batteries shifts loads to the utility automatically. Batteries cover 24-48 hours for most homes; longer storms lean on the grid or a generator. Winter output in the northern US runs 30-50% of summer peaks because of lower sun angle and shorter days, not cold. The Tesla Powerwall 3 stores 13.5 kWh, enough for most overnight gaps, and net metering catches the seasonal mismatch. Solar isn't a switch that flips off in bad weather. It's a faucet that turns down, and the rest of your system makes up the difference.
I logged a week of Solis 5G inverter data during a heavy December storm in northern England. The array (5.2 kW with LONGi panels) produced 2.8 kWh on the worst day and 14 kWh on the best, versus roughly 23 kWh average summer output. The drop is real, but it never hit zero outside of nighttime.
What Happens to Solar Output on a Cloudy Day?
Heavy overcast cuts irradiance from the 1,000 W/m2 reference to roughly 100-200 W/m2, so output drops to 10-25% of peak (NREL NSRDB). Light haze lets through 40-60%. Panels keep working because silicon absorbs diffuse light across the same wavelengths it catches in direct sun, just at lower intensity. Under 100 W/m2 cloud, a 400W panel generates roughly 40W, well above the inverter's startup threshold. In practice, a 6 kW system in Manchester makes around 1.5 kWh/day under heavy December overcast versus 35 kWh on a clear June day.
Diffuse light hits from every direction, so on a bright cloudy day the whole sky becomes the source, and tilt angle matters less, a horizontal panel can beat a steeply tilted one with no direct beam. The kicker: bright cumulus clouds can briefly push output to 110-115% of rated peak through edge-of-cloud reflections. The inverter clips it, but I've seen the spike on every monitoring system I've worked with in summer.
Does Solar Production Stop Completely Without Sun?
Production drops near zero only after sunset or under extreme conditions like a total eclipse. NREL's NSRDB records minimum daytime irradiance of 20-50 W/m2 during the worst storms, 1-3% of rated output, at which point the inverter sits below its DC minimum and shuts off. Eclipses are a separate case: the August 2017 North American eclipse cut California's fleet output by roughly 70% over 90 minutes, then snapped back, handled fine because operators saw it coming a decade ahead. In residential terms, "no sun" means nighttime or extreme storm overcast, both recoverable the moment photon flux returns.
How Do Batteries Cover Outages and Cloudy Weeks?
A 13.5 kWh Tesla Powerwall 3 covers roughly 24 hours of typical US household load (28-30 kWh/day per EIA) if you cut back heavy appliances. Divide usable capacity by hourly draw, allow ~90% round-trip efficiency, and you get bridge time. For a cloudy stretch it runs the other way: if overcast cuts production from 25 kWh to 5 kWh/day, you're 20 kWh short. One Powerwall covers most of a day; two handle 1.5-2 days; beyond that you pull from the grid or a generator. Not sure how many kWh you need? Our home battery sizing guide walks through the math.
| System | Storage capacity | Days of typical home backup | Best fit |
|---|---|---|---|
| Anker SOLIX aPower 2 | 2.0 kWh usable | 2-4 hours (partial loads) | Apartments, small cabins, portable solar backup |
| Tesla Powerwall 3 | 13.5 kWh usable | 0.5-1 day | Most US households, full home backup |
| Enphase IQ Battery 5P (x3 stack) | 15 kWh usable | 0.5-1 day | Modular sizing, AC-coupled systems |
| SolarEdge Home Battery 10 kWh | 9.7 kWh usable | 0.4-0.7 days | SolarEdge inverter owners |
| Powerwall 3 x 2 | 27 kWh usable | 1-2 days | Heavy users, EV charging, electric heat |
Is two days of storage enough for a Pacific Northwest winter? Honestly, no. Going fully off-grid in Seattle or Portland needs a much larger bank (40-60 kWh) or a generator; stacking five Powerwalls to ride out two annual storm weeks rarely pays. Net metering or a grid tie carries most of the load. For full grid independence, the off-grid solar packages guide breaks down real pricing from $15k to $45k.
What About Grid Backup When the Sun Disappears?
Grid-tied systems draw the shortfall from the utility automatically. When production drops below consumption, the meter runs the other way and you pull grid power, no switch to flip. Net metering does the seasonal averaging, with summer surplus credits offsetting winter shortfalls. California's NEM 3.0 cut those credits by roughly 75% in April 2023, pushing new installs toward batteries, but in full-retail net metering states (much of New York, Massachusetts, the Mountain West) the grid still works as a free virtual battery year-round.
The catch: a standard grid-tied inverter won't run during an outage even in sunshine. UL 1741 anti-islanding rules force it offline whenever the utility drops, to protect line workers. For power during outages you need a backup-capable battery inverter (Tesla Powerwall, Enphase IQ8, SolarEdge backup interface) or a transfer switch with generator. Too many homeowners learn this on the day of their first outage: "I have solar, why is my house dark?"
How Much Less Do Panels Produce in Winter?
Winter output in the mid-latitude US runs 30-50% of June peaks (EC JRC PVGIS). The cause is geometric, not thermal: shorter days, lower sun, worse angle of incidence on fixed-tilt south-facing panels. PVGIS estimates Chicago (41 deg N) at roughly 65 kWh per installed kW in December versus 145 in June (a 55% drop), Phoenix (33 deg N) at 110 versus 175 (37%), and Seattle (47 deg N) at 30 versus 150 in July, a brutal 80% spread.
Cold air actually helps efficiency. A typical TOPCon panel's temperature coefficient is around -0.30%/deg C, so at 5 deg C cell temperature it produces about 6% more rated output than at 25 deg C. Winter loses more from sun angle than it gains from cooler cells, but the cold itself is a benefit, covered in our piece on why panels work better in cold. Snow is separate: light dustings clear within hours on glass panels at 20-40 deg tilt, but heavy accumulation produces zero until it slides or melts, and automated clearance rarely pays. Modern panels survive 25 mm hailstones at 23 m/s (IEC 61215); the bigger concern is encapsulant micro-cracking from thermal cycling, which shows up as accelerated degradation.
How Do You Size a System for Low-Sun Periods?
You don't oversize for winter, you split the problem across array, batteries, and grid. Sizing the array for December consumption would need 2-3x the summer panel count, wasted nine months of the year through clipping. A practical setup uses three layers: the array meets annual average consumption, batteries handle overnight and short cloudy stretches, and net metering or grid backup absorbs the seasonal mismatch. Most US systems run at 80-100% of annual consumption. For heavy-cloud climates (Pacific Northwest, UK, northern Germany), a typical config:
- 7-9 kW array (summer surplus, winter partial)
- 13.5-27 kWh battery storage (overnight + short outage)
- Grid connection with whatever net metering exists
Going larger doesn't pay back. A fourth Powerwall to ride out the worst week usually returns below 4-5%, under what index funds do. For pulling more from existing panels, our guide on increasing solar PV yield by 20% covers orientation, optimizer placement, and dust management.
Summary
Solar panels keep generating in cloudy weather, typically 10-25% of peak under heavy overcast and 40-60% under light cloud. Grid-tied systems draw from the utility automatically when production falls short. Batteries cover overnight and short outages; longer stretches lean on net metering or a generator. Winter output drops 30-50% in mid-latitude regions from sun angle and day length, not cold. A sensibly sized system splits the problem across array, battery, and grid rather than oversizing one component. The Powerwall 3 covers 24 hours for most homes; two handle most multi-day storms. For seasonal strategy, the solar system optimization guide covers the full picture.