Solano County, California lost 23,830 Pacific Gas & Electric customers in a single event on May 3 — the most customers hit by any U.S. outage in May 2026. Southern California Edison spent the month restoring big ones of its own: 11,232 customers in Ventura County on May 25, a cluster of 9,200-customer events in Riverside County on May 2, and 8,968 in Los Angeles County on May 8.

A substation transformer failure in Ontario County, New York took out 11,407 Rochester Gas & Electric customers on May 26. Baltimore County, Maryland logged an outage on every single day of the month, and led the country in animal-caused outages while it was at it.

Which states had the most power outages in May 2026?

Massachusetts, New Jersey, and Maryland topped the country for outage density over the past 30 days, with Massachusetts averaging 774 outage events per county across its 14 counties.

StateEventsCounties HitEvents per CountyMassachusetts10,84214774New Jersey15,32621730Maryland10,84515723Connecticut4,75210475Alabama28,45860474Texas96,247225428California23,05857405

Texas had the largest raw count by a wide margin: 96,247 events across 225 counties. Its size pulls the per-county density below the Northeast corridor, where the same story keeps repeating — small, dense states where every county gets hit. The density-versus-volume tradeoff is the core decision in picking a sales territory.

New Jersey’s number is JCP&L and PSE&G splitting the state. Massachusetts is National Grid and Eversource territory.

Which counties had the most outages last month?

Harris County, Texas led the country with 17,743 outage events across four reporting utilities. Jefferson County, Alabama (Birmingham) was second at 5,406 from Alabama Power alone.

CountyStateEventsUtilitiesHarrisTexas17,7434JeffersonAlabama5,4061TarrantTexas4,4992NuecesTexas4,2651JacksonMissouri4,2542BexarTexas4,0142MontgomeryTexas3,9206DallasTexas3,6871Fort BendTexas3,4852Salt LakeUtah2,9732

The one to watch is Nueces County — Corpus Christi. 4,265 events from a single utility (AEP Texas) puts it fourth in the country, ahead of Bexar and Dallas, counties with five times its population or more. The ZIP-level breakdown below shows where inside the county those events landed.

The chronic list runs long: counties with at least one outage on every day of the month. Baltimore County, Maryland led again with 3,741 events across 31 days. Anne Arundel County, Maryland: 1,333 events, 31 days. Alameda County, California: 1,243. Bell County, Texas: 1,168. Allegheny County, Pennsylvania: 1,023. The 12-month version of that leaderboard shows which of these are permanent residents.

What were the biggest single power outages in May 2026?

The largest single May 2026 outage was Pacific Gas & Electric’s May 3 event in Solano County, California, which knocked out 23,830 customers. Power was back in about 90 minutes, and no cause was published.

The rest of May’s biggest events, ranked by customers hit:

• May 7, Northampton County, Pennsylvania. Met-Ed, 17,489 customers. No cause given.

• May 9, Contra Costa County, California. PG&E, 12,216 customers out for 14.3 hours. Cause: broken underground equipment.

• May 26, Ontario County, New York. Rochester Gas & Electric, 11,407 customers. Cause: substation transformer failure.

• May 25, Ventura County, California. Southern California Edison, 11,232 customers, restorations running 8 to 14 hours. Logged as “Equipment Problems.”

• May 7, Honolulu, Hawaii. Hawaiian Electric, 9,646 customers.

• May 2, Riverside County, California. SCE logged a string of near-simultaneous events around 9,220 customers each, with restorations from 8 to 19 hours. All of them sat in “Analyzing Problem” status throughout.

Equipment is the through-line. Broken underground equipment in Contra Costa, a failed substation transformer in Ontario County, “Equipment Problems” in Ventura. None of May’s biggest events came with a storm attached — the hardware just gave out.

Which ZIP codes had the most power outage activity in May 2026?

ZIP 77535 in Liberty County, Texas — the Dayton area, northeast of Houston — logged 805 outage events in 30 days, the most of any ZIP in the country.

The bigger pattern is the Texas Gulf Coast. Five Corpus Christi ZIPs sit in the national top 25:

• 78415 — 738 events

• 78418 — 579 events (Flour Bluff, Padre Island gateway)

• 78411 — 384 events

• 78412 — 355 events

• 78413 — 317 events

That’s 2,373 events across five adjacent ZIPs, all AEP Texas territory. Add 78382 in Rockport (528 events), 78332 in Alice (372), 78102 in Beeville (346), and 78566 in Cameron County (321), and the South Texas coastal corridor is the densest outage cluster in the country at ZIP level.

Outside Texas, 36207 and 36201 in Calhoun County, Alabama (Anniston) combined for 835 events, and 70466 and 70454 in Tangipahoa Parish, Louisiana combined for 671.

These counts come from binning each geolocated event to the ZIP its coordinates fall inside, which is what makes ZIP-level data worth pulling — the county number hides this much variation. For the history under any specific ZIP, see how to find power outages by ZIP.

What caused the May 2026 outages?

Most May outages had no public cause. “Pending Investigation” alone covered 33,443 events, and the full family of “Under Investigation,” “Currently Assessing,” “Unidentified,” and blank labels adds tens of thousands more. Why utilities don’t explain outages covers how routine this is.

Of the events that did get a cause:

• Equipment failure led on volume: “Equipment issue detected” (3,731 events), “Upgrading Equipment” (2,951 at 9.1 hours average), plus the usual long tail of equipment labels. In Texas, a single “Equipment Issue” label accounted for 296,000 customer interruptions on its own.

• Weather was steady: “Inclement Weather” (3,588 events), “Storm Damage” (2,697 at 7.6 hours average), “Weather Related” (2,117 averaging 82 customers per event).

• Tree contact logged 2,317 events at 7.2 hours average — fewer events than weather, but consistently among the slowest to fix.

• Planned work added roughly 19,000 events across the various maintenance labels. Filter these out when looking for pain; nobody buys a battery over an outage they got a postcard about.

What were the strangest outage causes in May 2026?

Cars hit utility poles more than 1,000 times in the past 30 days. The vehicle-accident labels combined for over 1,050 events averaging around 100 customers each, and one utility’s “CAR POLE” label averaged 356 customers per event with a peak of 3,856.

Animals took the grid down about 2,900 times. The most damaging label was “Animal in Equipment”: only 38 events, but averaging 619 customers and 11.1 hours each. Baltimore County, Maryland led the country in animal-caused outages with 105 events — the same county that logged an outage on every day of May. The squirrels are not helping.

The rest of May’s odd-cause ledger:

• Lightning. Roughly 700 events at about 8 hours average restoration.

• Pole fires. 98 events at 12.1 hours average, the slowest fix in this list. We covered why pole fires take so long to restore in a dedicated post.

• Dig-ins. 86 events of someone putting a shovel or excavator through a buried line.

• Wildfire de-energizations. 24 preemptive shutoffs averaging 238 customers each.

If you like this genre, the strangest outage causes in the data goes much deeper — balloons included.

The strongest door-to-door opener in energy sales right now is a homeowner naming their own bad week back to them. “We heard Pacific Power had an outage around here last Friday afternoon. A tree took down a line. Did you folks get hit?” It’s true, it’s fresh, and it never says the word solar. That’s why it works.

Most doors close in the first five seconds, and almost none of them close because of the words. They close because the homeowner pattern-matched you to “someone selling something” before you finished your sentence. A good opener breaks that match. A bad one confirms it. Here’s how the common ones stack up, ranked by how reliably each one earns the next thirty seconds.

What is the best door-to-door sales opener?

The fresh-outage observation is the best opener, and it isn’t close. You name a real event the homeowner lived through days ago, ask about their experience, and say nothing about what you install.

“Hi, I’m [Name] with [Company]. We heard Pacific Power had an outage around here last Friday, a tree took down a line. Did you folks get hit?”

It works on three levers at once. The specificity (a named utility, a day, a cause) reads as a neighbor who noticed, not a rep with a clipboard. The question is about them, not your product. And the pain is recent, so you’re not creating a problem in their head. You’re touching one that’s already there. For a backup-power rep, a fresh outage is what a hailstorm is to a roofer: the event that tells you which doors are worth knocking, and the reason the homeowner is willing to talk.

The one requirement is that it has to be true for that house. Which is the whole catch, and we’ll come back to it.

The opener tier list

S-tier: the fresh-outage observation

Covered above. The only opener that combines a real trigger, a question about the homeowner, and zero product mention. Its ceiling is the highest and its floor is still polite. The catch is sourcing: you need to actually know the house lost power, and when. Get that wrong and you’re a stranger making things up on a porch.

A-tier: the named neighbor

“I was just down the street working with your neighbor Adam on his place, and figured I’d stop by. Are you the homeowner?”

Social proof plus a decision-maker check in one breath. Naming a real neighbor borrows trust you haven’t earned yet, and the “are you the homeowner” tag means you don’t waste the pitch on a teenager or a renter. It loses to the outage opener on freshness. It gives the homeowner a reason to trust you, but not a reason to care today. It also wears out fast on a single street, since the fourth house knows you said the same thing at the first three.

A-tier: the curiosity pain point

“Quick question. Have you noticed your electric bill creeping up this year?”

Opens a loop instead of triggering the brush-off, and like the outage line it never mentions the product. The meter-change variant (”did you know they swapped your meter, and how that changed your credits?”) works the same way. It ranks just under the neighbor opener because the pain is real but not fresh. Everyone’s bill is up, so it lands as a general gripe rather than a thing that happened to them on Tuesday.

B-tier: the time-respecting easy out

“I’ll be quick. If I’m interrupting dinner I can be gone in ten seconds, fair?”

Giving the homeowner permission to dismiss you defuses the reflex to dismiss you. It buys breathing room, which is real, but it’s a frame, not a hook. On its own it earns you ten seconds and nothing to fill them with, so it works best bolted onto one of the openers above rather than carrying the door by itself.

C-tier: the generic question

“Hey, quick question for you...”

Better than pitching, because at least it asks instead of tells. But with nothing specific behind it, the homeowner’s guard stays up while they wait to find out what you’re really there for. Forgettable. It doesn’t lose you the door, but it doesn’t win you anything either.

F-tier: the feature-led pitch

“Hi, would you like to learn how solar can save you money?”

This is the door-closer. You’ve announced you’re selling, named the product, and asked them to volunteer for a pitch, all before saying one thing they care about. The homeowner doesn’t even hear it as a question. They hear “salesperson,” and the door is already moving. Every rep knows not to do this, and tired reps do it anyway around door ninety.

Why does a fresh outage make the best opener?

Because it’s the only opener where the homeowner already agrees with your premise before you knock. They sat in the dark. They lost the food in the fridge. They know the grid let them down, so you skip the part where you convince them there’s a problem and go straight to the conversation.

The data backs how often this trigger is sitting there waiting. Erie County, New York had outage events on 76 of the last 90 days. East Baton Rouge logged 74. These aren’t storm anomalies. They’re chronic-outage counties where some street is freshly dark almost every day of the month. And where restoration drags, the frustration runs deeper: Oregon averages 7.6 hours to restore power per event, and Nebraska 8.8. A homeowner who spent most of a workday without power is a different conversation than a cold knock. We get into what to say after the opener lands in a separate post.

Generators, batteries, and solar all sell on the same pain here. The outage is the trigger; the product is just which fix you carry.

What makes a door-to-door opener fail?

Three things kill an opener, and they’re the inverse of what makes the outage line work. Leading with the product tells the homeowner you’re selling before you’ve earned a reason to listen. Saying nothing specific leaves their guard up. And skipping the introduction entirely, diving into “we heard there was an outage” with no name and no company, trips the “who is this stranger” reflex even when the rest of the line is good.

That last one matters for the S-tier opener specifically. The outage observation is strong enough that reps want to lead with it cold, but a homeowner’s first question is always who are you. Bolt a quick identity line on the front (”Hi, I’m [Name] with [Company]”) and then drop the outage line. The introduction costs you two seconds and saves you the reflex.

One more failure mode is getting caught on the follow-up. When the outage opener works, the homeowner sometimes asks “wait, how did you know we lost power?” Have a clean, true answer ready: the utility reported it themselves. Every outage you’re working off of is one the power company posted on its own public outage map. “Your utility put it on their outage map, and I’m just knocking the street” is the whole truth, and it lands as ordinary instead of surveillance-y.

How do reps know which houses lost power?

The utility tells you. Every power company posts its outages on a public outage map, down to the area and cause, and the named-home version of that same data is what makes the S-tier opener usable instead of theoretical. The hard part is turning “this feeder lost power Friday” into “the Hendersons at 14 Oak Street lost power Friday,” a specific, knockable door.

That’s the product. GridProfile’s per-home outage leads resolve a fresh outage down to the owner name and street address for every home that lost power in your county, with the homes that already have solar, battery, or a generator dropped so you never knock a door that’s already equipped. You walk up already knowing the house went dark Friday, which is the one fact the best opener in the business depends on.

You don’t need the named file to use the opener. You can knock a street you know lost power and play the odds. But the named list is the difference between “we heard there was an outage around here” and knowing, before you raise your hand to knock, that this house is the one. Pricing scales with how tightly you filter: loosen the filters for more doors at a lower price each, tighten for a smaller premium set.

A couple of mechanics worth keeping regardless of which opener you run. Stand a step back from the door instead of crowding it, angle your stance, keep your hands visible. And when you set the next step, offer two times instead of asking an open question: “Should I swing back tomorrow at 6:30 or Saturday at 10?” beats “when works for you?” every time.

The opener earns you the conversation. The fresh outage is what earns you the opener, so grab a free grid profile to see which ZIPs in your territory went dark this week, or request your county for the named doors behind them.