Host Sold Two Four Seasons for $1.1B. The Per-Key Math Tells a Different Story.
Host Hotels sold 569 luxury keys for $1.93M each and called it capital recycling. The unlevered IRR looks clean at 11%... until you ask what replacement assets at that yield actually look like in 2026.
$1.1 billion for 569 keys. That's $1.93M per key across two Four Seasons properties (Orlando and Jackson Hole). Host is calling this capital recycling. Let's decompose what they actually did.
The stated unlevered IRR is 11.0%. The EBITDA multiple on exit came in more than 4 turns above Host's own trading multiple. On paper, this is textbook execution: sell assets where the private market values them higher than the public market values your stock, then redeploy into buybacks or acquisitions where the implied cap rate is more favorable. Host returned nearly $860M to shareholders in 2025 through repurchases and dividends. They've sold $5.2B and acquired $4.9B since 2018 while increasing Adjusted EBITDAre per key. The portfolio is getting smaller and (theoretically) more profitable per unit.
Here's what the headline doesn't tell you. The $500M taxable gain means roughly half the sale price was appreciation above basis. That's a strong exit. But the reinvestment problem is real. Host now needs to deploy that capital into assets generating comparable risk-adjusted returns in a market where luxury cap rates are compressed and construction costs have pushed replacement cost per key past $700K in most primary markets. Buying back stock at $19-20 (against analyst fair value estimates near $20.17) isn't exactly a screaming discount. The 35.26% one-year total shareholder return looks great in the rearview mirror. The question is what the next billion of deployed capital earns.
I audited a REIT once that executed a similar strategy... sold trophy assets at peak multiples, returned capital to shareholders, then spent two years sitting on dry powder because nothing penciled at the yields they'd promised investors. The stock drifted. The narrative shifted from "disciplined recyclers" to "can't find deals." Host's management team is sharper than most, but the math problem is the same. An 11% unlevered IRR is the benchmark they just set for themselves. Every future acquisition gets measured against it.
The condo residual ($17M recognized, $20-25M remaining) deserves a closer look. It suggests the Jackson Hole asset carried a residential component that contributed meaningful exit value beyond the hotel operations. Investors modeling Host's go-forward portfolio should strip that out when comparing per-key economics. The hotel-only implied price per key on that 125-room property is almost certainly north of $2M.
Here's what this actually means if you're an asset manager or owner evaluating your own hold/sell math right now. Host just demonstrated that the bid-ask spread between public and private luxury valuations is wide enough to drive a truck through. If you're sitting on a luxury or upper-upscale asset with significant appreciation above basis, get a current broker opinion of value this quarter. Not because you should sell... because you need to know what your capital is worth deployed elsewhere versus where it sits today. Run your own unlevered IRR from acquisition to a hypothetical disposition at today's private market pricing. If that number is north of 10% and your go-forward NOI growth assumption is sub-3%, you owe it to your investors to have the conversation. The window where private buyers pay these multiples isn't permanent. It never is.