The assumption doing the most work here is stabilized occupancy, and you already know that. But the second assumption, the one that will actually determine whether 8-cap is achievable or fictional, is who captures the care margin. Landlord-only versus landlord-operator is not a minor structural choice. If you lease to an operator, you're underwriting a triple-net or modified-gross rent that the operator can afford to pay while keeping their own EBITDA intact. That rent ceiling tends to land well below what the gross care revenue implies, because operators in a new licensing category carry real startup risk and will price that into what they'll commit to on a lease. Your $60k gross revenue number could support a very different rent depending on whether the operator is capitalized, experienced, and willing to sign a term that a lender will credit.
On the comp problem: you're right that there are no sales comps, but you can build a synthetic cap rate from the operator side. Get the proforma stabilized NOI to a number you can defend, then find what institutional or regional AL investors have been paying in adjacent markets for 8 to 16 bed licensed facilities. NIC tracks this at a macro level; your commercial broker may have access to more granular regional data. The cap rate you'd need at exit depends heavily on whether the asset is perceived as a real estate play or an operating business, and that perception shifts with licensing clarity. A buyer in 2027 looking at a stabilized, licensed facility with 18 months of occupancy history under a recognized zoning category is looking at a different risk profile than what you're staring at today.
The risk you haven't mentioned is licensing timeline slippage. Five pending applications means five operators learning simultaneously how the county processes these, and that process is almost always slower and more expensive than the planner's timeline implies. A gap between your carry cost assumption and actual time to first resident is where this underwriting can break.
A few things I'd want to know from you: what does your current model assume for months to stabilized occupancy, and are you treating licensing and buildout as sequential or parallel?