September 28, 2026 · By Phil

Which Costcos would make the best data centers?

Six hundred and seventeen warehouses, one dumb question.

Part 1: Initial Triage

Given all the community outrage about data centers lately, I've decided it's time our industry finally answered one of its oldest questions.

Which Costcos should we convert into data centers?

This mini-series will run through a mock site selection exercise to see which Costcos appear to have proximity to the best electrical grid infrastrucure. We'll call it Project Kirkland.

Project Kirkland Basics

Load: 50 MW at first energization, expandable to 150 MW.
Shell reuse: the warehouse stays. Roughly 150,000 square feet, high clear height, a slab rated for pallet jacks full of canned goods, huge parking lot.
Flood: outside a FEMA Special Flood Hazard Area.
Transmission: a 230kV substation within 2 miles.

Project Kirkland Fine Print

I've invented a client. Their requirements are made up. I want to be completely clear about that, because the entire point of this exercise is the line between what data tells you and what you decide.
It's just a fun exercise. Nothing in here should be taken as a recommendation.
There are some obvious reasons this exercise is silly:
- We're searching for proximity to transmission infrastructure and cross referencing with queue data, but we don't check to see if there is actual headroom at these nearby subs. The point of the initial screening tools on this site are just to make the initial kill list, not the eventual shortlist.
- Costcos are usually about 150,000sf, so as of today probably too small to make much use of the gigawatts we are typically searching for these days. But we do see local/regional inference plays in the 50-150MW range... and Costcos do have a huge parking lot for expansion, plus NVDA says they're targeting 1MW racks, so who knows!

Anyway, we may be loud, we may be power-hungry and water thirsty, but a Costco is a worse neighbor. Twelve hundred cars a day, a gas station, a tire bay... Let's see which ones we should take!

Building the list

There are about 650 Costco warehouses in the United States. I pulled them from OpenStreetMap, filtered to actual warehouses and cut to the lower 48. That leaves 617.

Then I ran all 617 through Site Scanner, which is the tool on this site that takes a coordinate and tells you what's around it: nearest substations and their voltages, the franchise utility, the flood zone, the watershed. Six hundred and seventeen scans, about six minutes.

The funnel

Stage	Sites remaining	Share of all Costcos
All Costco warehouses in the lower 48	617	100%
Outside a FEMA Special Flood Hazard Area	592	96%
↳ and any ≥100 kV substation within 10 mi	532	86%
↳ and ≥230 kV within 10 mi	292	47%
↳ and ≥345 kV within 10 mi	107	17%
↳ and ≥345 kV within 2 mi	41	7%

Look at the first two rows and sit with them for a second.

I expected flood plain and transmission proximity to cut the list way down. They eliminated just 14% of Costcos.

Ninety-six percent of Costco warehouses are outside a Special Flood Hazard Area. Ninety percent have a 100 kV or better substation within ten miles. Which, on reflection, is obvious: Costco spent decades and enormous sums picking sites near infrastructure, on dry, flat, commercially zoned land with good road access. They are, as a portfolio, extremely well sited. So the flood screen isn't weak, the population was pre-filtered by somebody else's very good real estate department.

Here's each screen applied on its own:

Screen	Costcos passing	Share
Outside a Special Flood Hazard Area	592	96%
≥100 kV substation within 10 mi	556	90%
≥230 kV within 10 mi	310	50%
≥345 kV within 10 mi	111	18%
≥500 kV within 10 mi	29	5%
≥230 kV within 2 mi	175	28%
≥345 kV within 2 mi	43	7%
≥500 kV within 2 mi	14	2%

Voltage alone at ten miles: 18% survive. Distance alone at two miles with a modest voltage floor: 28% survive. Put them together and you're at 7%.

So the screen that does the work isn't flood, and it isn't "near transmission." It's high voltage AND short distance, jointly. Ten miles from 345 kV is a different project than two miles from 345 kV. It's a transmission line build, a new right-of-way, possibly a new switching station, and years.

Here again, a 150,000sf building isn't very big, so there's really no reason we shouldn't be looking at 230kV substations or even smaller. Utilities can typically provide multiple smaller feeds from such subs as the data center grows. A typical 230kV substation can still get us over a gigawatt, for example, if we get multiple feeds. But given the results, we're going to run with the 345kV threshold for now to keep the list more manageable.

The 41

You can download the actual Site Scanner reports for these sites here.

#	City	State	Nearest ≥345 kV substation	Distance	Franchise utility	Utility type
1	Covington	WA	500 kV — COVINGTON	0.54 mi	PUGET SOUND ENERGY INC	INVESTOR OWNED
2	Allentown	PA	500 kV — WESCOSVILLE	0.56 mi	PPL ELECTRIC UTILITIES CORP	INVESTOR OWNED
3	Wilsonville	OR	500 kV — PEARL	0.71 mi	PORTLAND GENERAL ELECTRIC CO	INVESTOR OWNED
4	Kennesaw	GA	500 kV — BIG SHANTY	0.75 mi	GEORGIA POWER CO	INVESTOR OWNED
5	Eastvale	CA	500 kV — MIRA LOMA	0.94 mi	SOUTHERN CALIFORNIA EDISON CO	INVESTOR OWNED
6	Rancho Cucamonga	CA	500 kV — RANCHO VISTA	1.06 mi	SOUTHERN CALIFORNIA EDISON CO	INVESTOR OWNED
7	Tempe	AZ	500 kV — KYRENE	1.19 mi	CITY OF MESA - (AZ)	MUNICIPAL
8	Pottstown	PA	500 kV — LIMERICK (500KV SUB)	1.32 mi	PECO ENERGY CO	INVESTOR OWNED
9	Spokane	WA	500 kV — BELL	1.32 mi	AVISTA CORP	INVESTOR OWNED
10	Mt. Juliet	TN	500 kV — WILSON	1.61 mi	MIDDLE TENNESSEE E M C	COOPERATIVE
11	San Bernardino	CA	500 kV — SAN BERNARDINO	1.62 mi	SOUTHERN CALIFORNIA EDISON CO	INVESTOR OWNED
12	Tumwater	WA	500 kV — OLYMPIA	1.78 mi	PUGET SOUND ENERGY INC	INVESTOR OWNED
13	Victorville	CA	500 kV — VICTORVILLE	1.94 mi	SOUTHERN CALIFORNIA EDISON CO	INVESTOR OWNED
14	North Hollywood	CA	500 kV — STATION E (TOLUCA)	2 mi	LOS ANGELES DEPARTMENT OF WATER & POWER	MUNICIPAL
15	Eden Prairie	MN	345 kV — EDEN PRAIRIE	0.2 mi	NORTHERN STATES POWER CO - MINNESOTA	INVESTOR OWNED
16	Maplewood	MN	345 kV — KOHLMAN LAKE	0.33 mi	NORTHERN STATES POWER CO - MINNESOTA	INVESTOR OWNED
17	Everett	MA	345 kV — MYSTIC	0.35 mi	FITCHBURG GAS & ELEC LIGHT CO	INVESTOR OWNED
18	Brooklyn	NY	345 kV — GOWANUS	0.54 mi	CONSOLIDATED EDISON CO-NY INC	INVESTOR OWNED
19	Minneapolis	MN	345 kV — TERMINAL	0.67 mi	NORTHERN STATES POWER CO - MINNESOTA	INVESTOR OWNED
20	Plain City	OH	345 kV — UNKNOWN157686	0.67 mi	UNION RURAL ELECTRIC COOP, INC	COOPERATIVE
21	Houston	TX	345 kV — BELLAIRE	0.94 mi	CENTERPOINT ENERGY	INVESTOR OWNED
22	La Habra	CA	345 kV — COYOTE CREEK	0.98 mi	SOUTHERN CALIFORNIA EDISON CO	INVESTOR OWNED
23	Chicago	IL	345 kV — UNKNOWN124436	0.99 mi	COMMONWEALTH EDISON CO	INVESTOR OWNED
24	Bedford Park	IL	345 kV — BEDFORD PARK	1.06 mi	COMMONWEALTH EDISON CO	INVESTOR OWNED
25	Plano	TX	345 kV — ALLEN SWITCHING	1.06 mi	TEXAS-NEW MEXICO POWER CO	INVESTOR OWNED
26	San Antonio	TX	345 kV — CAGNON ROAD	1.11 mi	CITY OF SAN ANTONIO - (TX)	MUNICIPAL
27	Westbury	NY	345 kV — EAST GARDEN CITY	1.16 mi	LONG ISLAND POWER AUTHORITY	STATE
28	Albuquerque	NM	345 kV — SANDIA	1.18 mi	PUBLIC SERVICE CO OF NM	INVESTOR OWNED
29	La Vista	NE	345 kV — UNKNOWN131651	1.2 mi	OMAHA PUBLIC POWER DISTRICT	POLITICAL SUBDIVISION
30	Pleasant Prairie	WI	345 kV — BAIN	1.2 mi	WISCONSIN ELECTRIC POWER CO	INVESTOR OWNED
31	Yonkers	NY	345 kV — SPRAIN BROOK	1.25 mi	CONSOLIDATED EDISON CO-NY INC	INVESTOR OWNED
32	Houston	TX	345 kV — UNKNOWN302970	1.28 mi	CENTERPOINT ENERGY	INVESTOR OWNED
33	Bayonne	NJ	345 kV — BAYONNE SWITCHING STATION	1.42 mi	PUBLIC SERVICE ELEC & GAS CO	INVESTOR OWNED
34	South Windsor	CT	345 kV — TAP136499	1.45 mi	FARMINGTON RIVER POWER COMPANY	INVESTOR OWNED
35	Omaha	NE	345 kV — OPPD BLONDO STREET	1.6 mi	OMAHA PUBLIC POWER DISTRICT	POLITICAL SUBDIVISION
36	Arlington	TX	345 kV — UNKNOWN305459	1.68 mi	ONCOR ELECTRIC DELIVERY COMPANY LLC	INVESTOR OWNED
37	Lewisville	TX	345 kV — NORTHWEST CARROLLTON	1.77 mi	TEXAS-NEW MEXICO POWER CO	INVESTOR OWNED
38	Avon	MA	345 kV — STOUGHTON	1.81 mi	MASSACHUSETTS ELECTRIC CO	INVESTOR OWNED
39	Auburn Hills	MI	345 kV — UNKNOWN128507	1.86 mi	DTE ELECTRIC COMPANY	INVESTOR OWNED
40	Richmond	TX	345 kV — O'BRIEN	1.9 mi	CENTERPOINT ENERGY	INVESTOR OWNED
41	Loves Park	IL	345 kV — UNKNOWN122295	1.95 mi	COMMONWEALTH EDISON CO	INVESTOR OWNED

A few things I notice, none of which the tool told me:

California has 140 Costcos and six finalists. The largest store count in the country, a 4% conversion rate, and a queue and permitting environment that would make you think hard about all six. Texas actually leads the finalist list with seven.

Some of these are silly. Brooklyn, at 0.54 miles from the Gowanus 345 kV substation, is a Costco in New York City. You probably aren't building a 150 MW data center there. The data is correct and the answer is ridiculous, which is a useful reminder that a passing score on four criteria is not a site.

Mt. Juliet, Tennessee is served by a cooperative and sits 1.61 miles from a 500 kV substation named Wilson. A co-op with a 500 kV neighbor is an interesting conversation, and a completely different conversation than the one you'd have with Southern California Edison.

Eight of the 41 have substations with no name. UNKNOWN157686, TAP136499, and six others. More on that in a moment.

What Substation Scout did not do

This is the part we actually care about when using screening tools.

Substation Scout did not rank these 41. It did not tell me that two miles was the right radius, or that 345 kV was the right floor, or that flood zones matter more than utility type. It does not know what a good data center site is. It has no opinion about Brooklyn.

I picked every threshold in this post. I made up the client. I chose to treat a 500kV substation as better than a 345kV one without checking whether either has a single megawatt of available capacity — which, spoiler for part two, is the actual question.

The tool answered exactly one kind of question: what is near this coordinate, according to public data, and where did that data come from? Every number above traces to a published federal or ISO source, and the reports say so on their face.

The interpretation is yours. That's not a limitation I'm apologizing for, it's the ethos of this site. The moment tools like these start telling you which site is best, it's smuggling in somebody else's assumptions about load, tariffs, timelines and risk tolerance.

A footnote on the data, which is also the point

Those UNKNOWN157686 substations aren't a bug in my code. They're unnamed in the source data, and they're not rare. Eight of my 41 finalists have an unnamed or placeholder-named substation as their nearest high-voltage neighbor.
That matters more than it looks. An unnamed record still carries a voltage and a location, so it screens fine. But you can't match it to an interconnection queue entry by name, which means for those eight sites the next question in this series is unanswerable on a technicality.
The substation layer I'm using is a 2022 publisher snapshot in which individual records carry source dates as early as June 2015. Only 28 of the 617 sites have all eight of their nearest substations named.
Voltages are as-published and not field-verified.

So the honest caveat on every number in this post is: this is what the public record says, and the public record is a decade old in places and incomplete in others.

Which raises a question: why is the best public map of the American grid this old? We are in the middle of the largest load growth event in fifty years, every serious person agrees interconnection is the bottleneck, and the freely available substation inventory has 2015 source dates on it. Sweet.

That gap — between how fast we need to move and how legible the grid actually is — is why I'm building this site.

Next: I try to check what's already queued up at these 41 substations.

Site Scanner and the Queue Wire are on this site. Every figure in this post came out of them, and every source is cited in the reports they generate.