I want to share the key findings with you.
The Number That Stopped Me
I ran a comprehensive analysis of every industrial building 100,000 SF and above across five Western U.S. markets: Inland Empire, Los Angeles, Orange County, Phoenix, and Las Vegas. I filtered for buildings with 4,000 amps or greater of electrical service, the threshold where a building can realistically support meaningful automation, robotics, or advanced manufacturing.
Here is what the data showed.
552 buildings have 4,000 amps or more. That is 19.2% of the total 100,000+ SF inventory across all five markets. Roughly one in five.
115 buildings have 6,000 amps or more. That is 4.0%.
59 buildings have 8,000 amps or more. That is 2.1%.
To put those numbers in context: 4,000 amps runs a conveyor system and some AGVs. 8,000 amps runs what Amazon is building in Shreveport, the most automated warehouse in the company's history, with over 1,000 robots across every production area. If your automation roadmap looks anything like what the major 3PLs and manufacturers are deploying today, your universe of viable buildings across five major markets is measured in the dozens, not the hundreds.
In Orange County, there are 5 buildings with 8,000 amps or more. In all of Los Angeles County, there are 3. In Phoenix, 2.
These Are Not Normal Buildings
The 4,000A+ cohort is structurally different from the broader industrial market. The average heavy power building was built in 2012 versus 2005 for all 100,000+ SF buildings. It averages 462,000 SF versus 359,000 SF. Over 52% have 36 foot or taller clear heights compared to just 24% of the overall inventory. And 45% were built since 2020.
When I cross-tabulated vintage against ceiling height, the pattern was unmistakable. Pre-2000 buildings with heavy power almost all have sub-28 foot ceilings. From 2020 onward, 79% are 36 feet or taller. The market has bifurcated. Power without height is incomplete. Height without power is incomplete. Both together is where the premium lives, and the supply is thin.
Why This Is Happening Now
The U.S. power grid faces its most significant capacity challenge in a generation. Five year peak demand forecasts have increased from 38 GW in 2023 to 166 GW in 2025, a 4.4x revision in just two years. Data centers are the largest single driver, projected to more than double from 62 GW in 2025 to 134 GW by 2030. But every gigawatt consumed by a data center is a gigawatt that is not available for the manufacturer or 3PL down the street.
Former Google CEO Eric Schmidt told Congress the U.S. needs an additional 92 GW to support the AI revolution. That is roughly 60 new nuclear plants. In the last 30 years, only two have been built. The interconnection queue, the pipeline through which new generation connects to the grid, holds 2,300 GW of proposed projects, but only about 20% ever reach commercial operation. The median wait from application to energization now exceeds four years.
Companies that fail to verify power availability before committing to a site face delays measured in years and costs measured in billions. AEP Ohio imposed a 27 month moratorium on new connections after receiving 30 GW of requests. Northern Virginia faces 7 to 10 year waits. Chandler, Arizona, voted 7-0 to reject a $2.5 billion data center.
This is not a data center problem. It is a grid problem. And it is now your problem too.
What the Biggest Companies Are Already Doing
Texas Instruments chose Sherman, Texas, for a $60 billion semiconductor investment, citing ERCOT's deregulated grid and 100% renewable contracts. Micron committed $200 billion across six fabs, choosing upstate New York specifically for carbon-free hydro and nuclear power. Amazon chose Shreveport, Louisiana, for its most automated fulfillment center because Louisiana offers the cheapest industrial electricity in the country at 5.61 cents per kWh. Rivian shifted its R2 production to Normal, Illinois, to leverage existing power infrastructure rather than wait years for a new connection in Georgia.
The pattern is consistent: the largest capital allocation decisions in American manufacturing are being driven primarily by power access, grid reliability, and energy cost structure. Location, labor, and rent come second.
What Each Market Looks Like
The Inland Empire has the most heavy power inventory (379 buildings at 4,000A+), but it is spread across a massive geography and the outer markets (Adelanto, Beaumont, Desert Hot Springs) account for a disproportionate share of the highest amp buildings.
Phoenix looks strong on the surface with 42 buildings at 4,000A+, but 35 of those cluster in one amp band (6,000 to 7,999) in one corridor (Litchfield Park). A tenant needing 10,000+ amps in Phoenix has almost no options.
Orange County has the highest concentration rate (31.7% of 100K+ SF buildings have 4,000A+) but zero buildings with 40 foot clear heights and heavy power. If you need modern spec with heavy power in OC, it functionally does not exist.
Las Vegas has strong new construction with 30.1% concentration, but the dock ratios and building configurations suggest distribution product, not manufacturing.
Los Angeles skews old. A third of its heavy power buildings were built before 1990, with 24 foot ceilings and limited dock access.
Three Things You Should Do Now
First, audit your power before your next lease. Your power needs will be 5 to 10x what they were a decade ago if you are automating. Call the utility before the broker. Confirm substation capacity, interconnection timelines, and available feeders at every site.
Second, target markets with power today. TVA territory, central Iowa, central Washington, and Sacramento's SMUD district still have headroom. California's investor-owned utilities (PG&E, SCE) are 2.7x the national average at 21.60 cents per kWh. Municipal utilities in the same state can be half that.
Third, evaluate behind-the-meter generation. 30% of planned U.S. data center capacity now includes on-site power, up from effectively zero a year ago. CHP systems, solar plus storage, and microgrids get you operational years ahead of utility queue competitors. The global microgrid market is projected to reach $228 billion by 2035.
I built a full research paper covering grid capacity, industrial electricity rates, interconnection timelines, and utility expansion plans across six U.S. regions, along with 10 real-world case studies from Texas Instruments to Amazon to Rivian. I can have it in your inbox this week if it would be useful. Just reply to this email and I will send it over.
Justin