How does a preferred equity sleeve actually sit in the capital stack on a 20 MW build-to-suit

The "protected position" language on preferred equity in a capital stack that thin, $60M of equity cushion on a $240M deal, is basically marketing. Northern Virginia data center deals have gone sideways on entitlements and power delivery timelines, not occupancy, so the risk profile is different from what you'd price in traditional real estate.

The PPA collateral on the second lien note is the more interesting asset here, especially if the offtaker is investment-grade. A hyperscaler PPA in that Gainesville/Prince William corridor has real transferable value because power is the actual scarce resource right now, not the building. 200 bps does not compensate you for sitting behind $180M versus sitting behind $60M with a lien on contracted revenue.

Push the smaller sponsor hard on who the offtaker is and what the cure and assignment provisions look like in the PPA before you decide cleaner docs aren't just cleaner because there's less to hide.

The 200 bps gap is probably the least of the problem here. What I'd be pushing on is that $60M senior on the 8 MW asset implies a loan-to-value that is actually pretty disciplined if the PPA counterparty is investment grade, whereas the preferred equity sleeve on the build-to-suit has you sitting behind 75 cents of every dollar in the stack before construction is even finished in Gainesville. Build-to-suit is the worst place to be in preferred equity because your "protected position" only means something after the senior lender is made whole, and on a 20 MW shell with no stabilized cash flow yet, there is no floor under that. I did a TC review on a similar structured deal in Ashburn two years ago where the sponsor used the exact same 14% preferred language and the waterfall language in the actual operating agreement had a pay-if-available carve-out that basically let them defer the preferred return if NOI didn't hit a threshold. Read the operating agreement, not the deck. The second lien note is a simpler instrument and simpler instruments are easier to enforce when things go sideways.

The one thing I have never seen work is a sponsor calling pref equity "protected" when the senior covers 75% of the stack. What actually happens is that word disappears from the conversation the moment there is a covenant breach on the construction loan and the senior lender controls the timeline. On the 20 MW build-to-suit, your $200k is sitting in a roughly $60M mezzanine slice, and if lease-up slips even six months, that 14% accrues but you are negotiating with a senior lender who has zero incentive to move fast for you. The second lien against the 8 MW with the PPA attached is a different animal, because I have seen power purchase agreements from AWS-tier offtakers anchor workouts in ways that a general real estate lien just does not, the offtaker keeps paying even when the sponsor is in trouble. Two hundred bps does not cover the difference between having a lien on a cash-flowing contract and having an equity claim behind $180M, and the cleaner documents on the smaller deal are not a coincidence, that is usually the sponsor who got burned once and fixed the paperwork.

The 14% looks better until you ask what the senior loan-to-cost actually is. $180M on a $240M stack is 75% LTC, which means your preferred tranche has maybe $60M of total equity cushion and you are absorbing first-loss on all of it before common takes a hit. "Protected" in sponsor decks usually just means ahead of common, not that you have real collateral.

The PPA assignment on the second lien is what I would be pulling apart right now. Who is the offtaker, how many years are left on that contract, and does the lien actually attach to the revenue stream or just the paper?

14% pref on a stack that thin is running closer to 11.5% net in Northern Virginia once you strip promote and fees.

"Protected position" is marketing, not legal standing, you are unsecured equity behind $180M.

What is the LTV on the senior alone, and who is the construction lender?

The $60M first on the 8 MW asset is the number I'd pull apart before anything else. If that facility was capitalized at, say, $72M to $78M total and the first lien is already at $60M, you are looking at maybe $12M to $18M of cushion for your second lien position in a liquidation scenario, and that cushion has to absorb transaction costs, workout fees, and any impairment to the PPA before you see a dollar. PPA assignment in bankruptcy is not clean either. A lot of hyperscaler leases have termination-for-convenience clauses, and if the tenant exercises that in a distress event, the collateral package you are counting on as a floor just lost most of its value.

On the build-to-suit side, 75% LTV at senior alone before preferred equity even appears is the actual math on that $240M stack, and a 20 MW ground-up in Northern Virginia right now has real cost overrun exposure given where steel and electrical gear lead times are sitting. The 200 bps yield gap is almost secondary to what I would want to know first, which is whether that build-to-suit lease is fully executed or still at LOI stage. An executed 15-year NNN with a creditworthy cloud tenant at that scale changes the risk conversation entirely. A signed LOI with a colo operator does not, and I have seen sponsors call both the same thing in a deck.

What never works is assuming a PPA assignment survives a receivership cleanly. In every distressed data center situation I have read about, the hyperscaler tenant has termination rights or step-in rights that get exercised before a junior creditor can monetize that collateral, so the floor on the second lien disappears exactly when you need it.

That said, $180M of senior against a $240M stack on a build-to-suit that hasn't stabilized yet means your preferred equity cushion is basically the sponsor's equity, which on a 20 MW build-to-suit in NoVA could get eaten by construction overruns alone in 2024 dollars.

Two hundred bps doesn't move me given that both positions have a real scenario where you're waiting out a workout at 36 to 60 months with no distributions.