The Chipmaker Went Shopping for Dirt

🏗 From The Dirt Up

Real Estate Insights. Success Mindset. One Step at a Time.

Issue #74 | October 25, 2026

Everybody watches the chips. The smart money watches the dirt. This week, the biggest chip company in the world showed you why.

🧠 Market Insights

The Chipmaker Went Shopping for Dirt

On Tuesday, SB Energy posted a short notice for its neighbors in Pike County, Ohio. Clearing and grading for phase one of the PORTS-Pike Technology Campus starts in the coming weeks. Doorhangers went up. Crews may need to blast when they hit rock.

That's a small notice. It's also the clearest sign yet of where this market is going.

Back in August, Nvidia said it had "secured land, power and shell (LPS) capacity" at that campus. Read that again. The company selling the most valuable chips on earth now treats land, power and shell like supply it has to lock up.

Here's the deal. SB Energy builds, owns and operates the campus. OpenAI signed a 20-year lease. Nvidia backs the land, power and shell behind the first 4.25 gigawatts of IT capacity. Its exposure is capped at $105 billion, and it only pays if OpenAI doesn't.

Translation: a chipmaker is guaranteeing the dirt so its chips have somewhere to go.

Jensen Huang saw this coming. In March, he told Stratechery's Ben Thompson that a one-gigawatt AI factory costs "probably $50, $60 billion," and about "$15, $17 or so, is infrastructure: land, power, and shell." What he wants "more than anything is that all the land, power, and shell would just get stood up faster."

Chips ship in months. Land, power, and shell take years. That's the bottleneck, and value flows to the bottleneck.

The Dirt Nobody Wanted

Look at where Nvidia went. PORTS-Pike sits on the decommissioned Portsmouth Gaseous Diffusion Plant, a former federal uranium site. DOE is still deactivating the old process buildings today.

It isn't pretty dirt. It's complicated dirt. But it has what pretty dirt usually lacks: a history of heavy power, industrial use, and a community that remembers the jobs. The first 800 megawatt phase is slated to come online starting in 2028. The plan calls for about 10 gigawatts of new generation, roughly 9.2 of it natural gas.

Here's the part landowners should circle. DOE leased 189 acres of the old plant to the project. The federal permitting dashboard puts the first data center complex on about 1,300 acres of private land next door.

The old plant got the headlines. Most of the acreage came from private land.

It's the same play everywhere. On Monday, AEP, X-energy and Amazon picked a retired Indiana coal plant site to study a 960-megawatt nuclear plant. In Alton, Illinois, a developer has a contingent deal on a closed 381-acre steel mill for a 200-megawatt data center. Every one of those deals starts with a site somebody else wrote off.

Land Alone Isn't the Moat

Here's the catch. Land comes first, but land alone doesn't win.

Amazon paid $427 million for George Washington University's Ashburn campus, in the heart of Data Center Alley. This week it withdrew its data center plans, agreed to a rezoning for housing and mixed use, and is donating about 25 acres to Loudoun County. If Amazon can't get a data center approved in Ashburn, "near the cluster" isn't a pitch anymore.

On Thursday, Business Insider covered a Bernstein study that scored announced North American data center projects on three questions. Has construction started? Is power available? Has the developer built before? The answer: about 170 of nearly 500 announced gigawatts will likely get built. Roughly a third.

Bernstein put it plainly: "Press releases are cheap…concrete is not."

Apollo's chief economist, Torsten Slok, made the same point from the money side on Wednesday. If the buildout gets delayed, his list of winners includes "holders of powered land and grid access," because "sites with secured power, interconnection and permits are the scarcest input."

Notice what he didn't say. He didn't say land. He said powered land with interconnection and permits.

The Money Is Getting Choosy

The big buyers are still spending. Meta narrowed its 2026 capital spending guide to $135 billion to $145 billion on Wednesday. Third-quarter capex hit $34.91 billion, more than its operating cash flow. It also created a new class of stock to help fund the growth.

Amazon has the money and is still waiting on power. Business Insider reported that AWS had Nvidia racks sitting in warehouses because some sites weren't ready to turn them on.

The leveraged buyers are a different story. CNBC reported CoreWeave's default insurance near 1,067 basis points, with the high-yield market "effectively on hold until after the midterm elections." Bond investors want permitting and regulatory risk settled before they lend.

Put it together. The investment-grade buyers are still coming. The junk-bond buyers are waiting on permits. Both want the same thing: a site where the land, the power, and the permission are already real.

✨ The Alpha Effect

Be the Bottleneck

Most people chase what's hot. Very few own what's scarce.

The crowd runs toward whatever moves fastest. Fast gets the headlines. But fast things get copied, and copied things get cheap.

The things that win over time are slow to build and hard to replace. Nobody can rush them. So the people who already have them set the terms.

That's true in land. It's true in a career. It's true in a reputation.

Think about the people you trust most in your business. You didn't pick them because they were quick. You picked them because they were hard to replace. They did the boring work for years before anyone noticed. Then one day everybody needed what they had.

Here's the trap. When the market gets noisy, it's tempting to trade your position for a headline. Sell the slow thing you own to buy the fast thing everyone's talking about. Jump because standing still feels like losing.

Standing still isn't losing if you're getting ready while you stand there.

The alpha mindset zigs while the crowd zags. The question isn't "what's hot?" It's "what's hard?" What takes years? What can't be faked with a press release? Do that work quietly while everyone else is busy announcing.

You don't have to be first. You don't have to be fast. You have to be ready when the fast money shows up and finds out it can't go around you.

Own the hard thing. Let the market come to you.

🚀 One Step at a Time

Action Steps This Week

1. Build your one-page "land, power, shell" sheet.

Keep it honest and plain. Land: acres you control, what's buildable, and how it's zoned. Power: your utility or co-op and the nearest substation or line you know of. Shell: any existing buildings, pads, rail or roads. Mark each item done, planned or just studied. When a serious buyer calls, you'll be the easiest site on their list.

2. Check your county's calendar before you market.

Call the planning office or read the next agenda. Is a data center pause or new ordinance coming? Does your zoning allow one at all? Counties all over the country are writing these rules right now. Finding out early beats finding out from a buyer's retrade.

3. Vet your buyer the way Bernstein does.

Ask three questions. Have you built a data center before? Where are you under construction today? Where does the power come from, and who's the end tenant? A buyer with real answers is worth more than a bigger offer backed by a press release.

Small steps compound. The chips will keep getting faster. The dirt won't.

Here's to building your success, one step at a time...

from the dirt up.

Kris "Dirt Dawg" McGee 🐾

The Dirt Dawg Den