Sizing data center exposure when two tenants are 40% of the rent roll

Question

I've been comparing the two ways I can actually get at this as a passive holder. Listed REITs, where I can read the 10-K, and a non-traded private fund a broker sent me at 1.25% management plus a 20% promote over an 8% preferred.

What's stopping me on the listed side is concentration. On one of them the top two tenants are north of 40% of annualized rent, and both are companies whose data center spend is currently a function of an AI capex cycle. That's not a diversified rent roll, it's a bet on two balance sheets continuing to want capacity at renewal. The private fund is worse on transparency and I'd be locked up 7 to 10 years.

If the sector ranks first for investment prospects and the only real debate is whether it's a bubble, how do you size a position where the demand story and the concentration risk are the same fact?

Rena Expert · Accepted Answer

Concentration and quality are the same fact here for a structural reason. The hyperscale tenants are the ones who can sign a 15 year lease at scale and actually pay it, so a rent roll skewed to two or three of them is a sign the landlord won the leases people wanted to win. What you're paid to worry about is renewal, not default. Those tenants have investment grade credit and will keep paying through a capex slowdown. The question is what rent they sign at in year 12 and whether they've built their own capacity by then.

So I'd underwrite lease expiry schedule and mark-to-market rather than tenant count. Pull the expiration table by year and ask what share of rent rolls in the next five. If in-place rents sit below current market and power is constrained in those markets, expiry is an upside event. If in-place rents were signed at the top of a leasing spike, expiry is where the bubble question actually shows up in your cash flow.

On the fund, 1.25% and a 20% promote over an 8% preferred is a market-ish structure, but the fee load only makes sense if the sponsor is doing something you cannot buy in the listed market, which usually means development or powered shell delivery rather than stabilized assets. If they're buying stabilized leased facilities, you're paying a promote for beta you can get at daily liquidity. Ask what percentage of the portfolio will be pre-leased at close.

One thing to check that people skip: the listed vehicles carry meaningful development pipelines funded with debt and equity issuance. If capital markets tighten while a pipeline is mid-build, the equity gets diluted at a bad price, and that risk has nothing to do with tenant credit.

ehernandez74 · Answer

I've been sitting on this same decision for a while and the lockup is what keeps deciding it for me. A 7 to 10 year lock on a sector where the main open question is whether current capital levels overshoot means I lose the one option that matters, which is being able to change my mind in year three.

tashaford30 · Answer

On the expiry table point, I looked at one and the weighted average remaining lease term was over 7 years with under 5% of rent expiring in the next two. That's a very different risk than I assumed from reading about the tenant concentration alone.

Sizing data center exposure when two tenants are 40% of the rent roll

3 replies