Open in interactive viewer → charts, metric popovers & call review

Blackstone's Data Center REIT Lands Amazon — and the Market Yawns

BXDC's inaugural deal beats the math it took to IPO, but a 15% drawdown and a 49% leverage print say investors wanted a different asset
BXDC · Earnings Call · 2026-11-16

A quarter of cash, then a $5 billion handshake

BXDC's third quarter is almost an accounting artifact: a REIT with no real estate. GAAP net income of $0.15 a share, FFO of $0.15, AFFO of $0.17 — earned on “cash and cash equivalents [that] generated $18.9 million of interest income during the quarter” — Anthony Marone, Chief Financial Officer · 2026-11-16. Total assets of $2 billion, GAAP book value of $19.34 a share, and zero debt. Management said the same thing three different ways: these results mean nothing yet.

Then the real announcement landed, filed the same morning as an 8-K. A signed definitive agreement for a 75% interest in a $5 billion portfolio — "13 powered shell data center assets across 2 campuses," fully leased to Amazon on a 12-year triple-net initial term and 32 years fully extended, with 2.75% annual escalators, a 7% going-in cap rate and an 8.1% GAAP yield. Amazon keeps the other 25%. The joint venture is the novel part: the counterparty is the tenant is the co-owner, on a full parent guarantee through the lease term. This is the inaugural transaction, and it was not auctioned — “This was a transaction that we negotiated directly with Amazon” — Nicholas Pell, President and Chief Executive Officer · 2026-11-16.

The strategic claim riding on it is large:

With CapEx expected to eclipse $800 billion in 2026 alone and to climb further in 2027, hyperscalers need access to a broad set of capital sources... we see the hyperscaler sale-leaseback market as an additional avenue of growth for us, one that doubles our total addressable market to more than $2 trillion by 2030.

Nicholas Pell, President and Chief Executive Officer · 2026-11-16

The buy box moved, and nobody wants to say so plainly

Last quarter's keyword set was dominated by process words. Buy box ranked third; blind pool was still in the vocabulary, because it still was one. Nick Pell then: “we're very focused on delivering our IPO proceeds into the specific buy box we laid out” — Nicholas Pell, President and Chief Executive Officer · 2026-08-04. And: “we're trying to prioritize for our IPO deployment into deals and assets that can pay rent upon closing” — Nicholas Pell, President and Chief Executive Officer · 2026-08-04.

The deal does pay rent at closing. But the box itself was redrawn. At IPO the pitch was turnkey data centers in Tier 1 markets at a low-6% cap rate. What they actually bought is a powered shell — land and building only, with Amazon owning and maintaining every server, rack and cooling system — at 7%, with "more than 2/3 concentrated in Northern Indiana" and the remainder in Mississippi. Mike Forman, in the closest thing to a confession on the tape:

When we created this company, we thought we were going to buy turnkey assets at a low 6% cap rate. And instead, we're buying powered shells at a 7% cap rate. We think that's very compelling relative value.

Mike Forman, Chief Investment Officer · 2026-11-16
The arithmetic is genuinely supportive — a basis "more than 50% below estimated turnkey replacement costs," rents "more than 70% below" comparable turnkey assets, an investment-grade tenant. Forman on the geography: “Indiana is like Chicago but better from a data center perspective” — Mike Forman, Chief Investment Officer · 2026-11-16. But note what has happened to the language. Tier 1 market is now the company's single top keyword; a quarter ago it was absent. When a company starts redefining its own category in the same quarter it makes its first purchase, that is the tell — the buy box didn't get abandoned, it got enlarged, and the stock's job is to decide whether that is discipline or drift.

The leverage line, and the cost of the pivot

The CFO did not bury the mechanics. “We expect the implied leverage level for BXDC following this transaction to be approximately 49%, but we remain committed to our long-term target of 40%.” — Anthony Marone, Chief Financial Officer · 2026-11-16 Financing is penciled at SOFR plus mid-to-upper 100s — “Slightly higher financing costs than during the IPO, but the yield on the asset, we think more than offsets that.” — Anthony Marone, Chief Financial Officer · 2026-11-16 So BXDC stretched nine points above target to deploy its full $1.9 billion of IPO cash, and still clears only a low-to-mid 5% AFFO yield against the mid-to-high 4% illustrative figure it showed on the roadshow. That is a real beat versus its own promise — but delivered by leverage and asset mix, not by buying better real estate than advertised.

The stock disagrees, and the tape explains why

BXDC is down 15.4% from its 2026-08-06 first-day close of $19.79 — trading at roughly 85 cents on a $19.34 book that is essentially all cash. Simultaneously, the private market just cleared a landmark, investment-grade, top-of-range deal with the most creditworthy tenant on earth. Those two facts have to be reconciled, and the answer is mostly beta. Data-center and AI-infrastructure equities have been sold hard across every window: traditional data center names down roughly 41% over 90 days, data center capacity down about 20% over 30, co packaged optics and High Bandwidth Memory complexes similarly weak. Global keyword flow is saturated — AMAT on fab equipment demand, CSCO on AI defense, CEZ and GRID.TO and VUZI all on data centers — and the tape is voting against the theme anyway. BXDC is a small-cap REIT whose only asset is cash and a signed contract, so it has no earnings cushion to absorb a sector de-rating.

The counterweight is policy. Forman: “following the election, I expect that it will be more difficult to develop than it was beforehand” — Mike Forman, Chief Investment Officer · 2026-11-16. Fewer new builds mean supply constraints tighten, which makes already-powered, already-leased assets more valuable — and makes the recycle capital motive among hyperscalers stronger, not weaker. That is a coherent thesis the public market is not yet paying for.

The setup into 1Q27

Forman last quarter: “We have more opportunities than we have capital for right now, frankly, which is a good place to be.” — Mike Forman, Chief Investment Officer · 2026-08-04 Still the framing. Pell is tracking "over $300 billion of newly constructed developer and hyperscaler-owned data centers expected to deliver" in the U.S. in 2027 alone, more than half of it outside historical Tier 1 — the mechanical reason he can claim an addressable market that doubles by 2030.

Timing: closing expected before year end, funded with $1.9 billion of cash plus an allocable $1.9 billion of asset-level financing. Fourth-quarter results will look like the third until it closes, and any dividend declared is a REIT tax requirement, not a run-rate proxy. 1Q27 is the first honest print. The bear case writes itself — an externally managed small-cap paying above its own leverage target for thinner assets in frontier markets, into a tape that hates the sector. The bull case is a 15% discount to a cash book and a deal template with the largest balance sheet in the world. The next twelve months decide which one this was.