Data-center-adjacent demand is doing a lot of work in the 2026 absorption forecast. How much of it do you underwrite?

The question I'd want answered before either choice is what happens to the improvements. If a staging tenant pours a heavy slab and brings in 2,000 amps and then leaves, you own that. It's not tenant-specific in the way a dental buildout is. So underwriting at generic rent but giving credit to the residual power and yard seems like a middle path. That's a slower way to lose fewer deals, which is roughly my whole approach to everything.

Losing deals for six months while your assumption holds isn't the same as being slow. Slow is when you can't tell which assumption you're defending.

On the capital side the version of this I've watched get funded is a hybrid. Base case at market logistics rent, then a separately identified upside case at the specialty rent, with the sponsor's promote structured to only really pay on the upside case. That way the LPs are buying the boring building and the sponsor is the one carrying the data-center bet. If the person bidding against you is underwriting the premium into base case and they're using other people's money, they may just have a different risk seat than you, not better information.

Small multifamily version of the same problem is a unit that a corporate tenant pays 30 percent over market for. I stopped underwriting those premiums after the second time one left. What I do now is take the premium as amortization of whatever I spent to serve them and nothing more. thicket, if you can get the specialty tenant to pay for the power upgrade over the term, you've converted the whole question into a capital recovery question instead of a rent question.

How would you even verify a staging tenant's story? On the residential side I'd pull employment and payment history. Is there an equivalent for a subcontractor whose entire demand comes from one campus decision made by somebody else?

wren, you look at their contract backlog and who the general contractor is, and in a lot of cases you get very little, which is the answer. On the land side I've watched this exact demand appear as ground lease inquiries for outside storage near substations, and the inquiries move around by the quarter. The land underneath doesn't care, which is why I like it. A building with 40 foot clear and heavy power cares a lot. thicket, I'd keep underwriting generic and stop counting the lost deals as losses, since you don't get to see how the winners' second five years go.