Most writing on hotel REITs takes apart what the investment corporations already own — how acquisition prices per room have polarised, how portfolio granularity differs, how far current appraisal values have drifted from purchase prices. All of it looks inside the 302 properties already held. This article inverts that view. Its subject is the size of the universe that no one has acquired yet. How many properties and how many rooms in Japan are not held by the seven listed REITs whose primary investment target is hotels — and what shape does that pool take when cut by property attributes: size, vintage and area? Counting the outline of that pool is what this article contributes.

Let us draw the line against our earlier work up front. First, the articles covering acquisition-price polarisation, portfolio granularity and the gap between purchase price and appraisal value were all breakdowns of held portfolios; their denominator was the properties REITs own. The denominator here is the properties REITs do not own. Second, the earlier mapping of acquisition headroom in areas REITs have not entered selected areas by demand signal — supply-demand tightness. This article defines the universe by property attributes (size, vintage, ownership), not demand; a different axis entirely. Third, where earlier work on the future supply pipeline counted rooms yet to be built, this article counts rooms that already exist and have not yet passed into institutional hands.

Conclusion — listed hotel REITs have touched only one-twentieth of the investable stock

Japan has 5,128 properties and 1,015,610 rooms with 100 or more guest rooms for which sales were actually observed online for near-term check-in dates. Of these, the seven listed REITs whose primary investment target is hotels hold just 194 properties and 45,273 rooms — 3.8% on a property basis, 4.5% on a room basis. Put the other way round, 4,934 properties and 970,337 rooms remain outside the listed hotel REITs.

To grasp what that means, set it against the seven REITs' combined acquisition spend. Since inception they have assembled 302 properties for roughly ¥1.84 trillion. Applying, by size band, the median price per room the REITs have themselves actually paid to the 970,337 rooms not held gives a theoretical pool value of about ¥23.7 trillion. Even applying a single median drawn only from the 100+-room properties of the three limited-service-focused REITs (about ¥19.24 million per room) yields roughly ¥18.7 trillion. The listed hotel REIT sector has left untouched a pool more than ten times the size of the assets it has taken in.

The acquisition of Hyatt Regency Tokyo (ハイアット リージェンシー 東京; opened September 1980, 712 rooms after renovation) by Japan Hotel REIT (ジャパン・ホテル・リート投資法人), announced in February 2026 at ¥126 billion, was the largest hotel purchase ever made by a domestic REIT. Even so, that single deal accounts for 0.07% of the rooms in the not-held pool and about 0.5% of its theoretical value. The pool is not of a size that any one large transaction visibly shrinks. We analysed the economics and operating performance of that deal in detail in JHR's ¥126 Billion Hyatt Regency Tokyo Acquisition — Operating Performance of the Largest REIT Deal on Record.

Cut by size — 75% of not-held rooms sit in the 100–299-room mid-size band

Breaking the universe down by room-count band shows clearly where the not-held stock piles up. The three mid-size bands — 100–149, 150–199 and 200–299 rooms — account on their own for 4,425 properties and 733,470 rooms, 75.6% of all not-held rooms. REIT penetration in these bands runs at 2.8–4.0% on a room basis, below the 4.5% overall average.

By contrast, penetration reaches 8.4% in the 300–499-room band and 6.8% above 1,000 rooms: the larger the property, the deeper listed REITs have gone. That is the rational outcome of how investment corporations behave. The bigger the ticket, the more thinly due-diligence costs and the fixed costs of operator arrangements can be spread, so large assets were taken in first. Read the other way, the 100–299-room band, where a deal runs ¥2–5 billion, falls below the standard ticket size of listed REITs, and most of that pool has therefore been left as it is. This band is the centre of gravity for portfolio build-out.

Cut by vintage — the not-held stock is new, not old

The phrase not-held pool invites a picture of ageing stock that listed REITs have avoided. The data rejects that intuition. By vintage band, room-basis penetration by the listed seven is highest for properties opened in the 1990s (8.8%), followed by pre-1989 (6.0%). It is lowest for the 2010s (2.6%) and 2020 onwards (2.1%).

In other words, listed hotel REIT portfolios are concentrated in large properties built around the bubble era, and have barely reached the newer limited-service stock delivered in volume during the inbound boom. Properties opened from 2010 onwards total 2,389 properties and 466,541 rooms, of which 455,580 rooms are not held — 47.0% of the entire not-held pool. About half of the pool that has not yet been acquired consists of relatively new buildings opened since 2010, and that is the finding with the most practical consequence in this article.

That structure changes how large-scale renovation risk should be read. In the conventional discussion, which assumed ageing stock, taking in the not-held pool looked synonymous with shouldering heavy capex. In reality the core of the pool is newer buildings at or around their first equipment-renewal cycle, so near-term capital expenditure is comparatively light. The 186,503 not-held rooms opened before 1989 are a different matter: at 37 years or more since opening, decisions on renewal investment versus rebuilding or conversion are approaching, and they should be treated as a separate class of stock. We examine where that older band splits between renovation and rebuilding, using data on 588 buildings, in 236,000 Rooms Under the Old Seismic Code, 19.1% of All Rooms — Renovate or Rebuild: the Dividing Line Across 588 Buildings.

Cut by segment — 630,000 rooms in limited-service, listed REITs reaching just 3.3%

By segment, the principal component of the not-held pool becomes clear. Limited-service properties (the business-hotel classification) account for 3,517 properties and 652,650 rooms, 64.3% of universe rooms, with penetration by the listed seven at just 3.3%. Full-service properties (the city-hotel classification) number 732 properties and 181,344 rooms at 6.9% penetration, and the deluxe classification 109 properties and 34,446 rooms at 11.0% — the higher the rate segment, the deeper the penetration.

That distribution is consistent with how investment corporations run their strategies. Full-service assets carry a larger ticket and offer more upside under variable-rent schemes. Limited-service assets are smaller per deal and require deal count to be stacked, so building a portfolio takes time. Despite the stability advantage limited-service properties offer on occupancy, this is the segment whose absorption lags furthest behind the size of its pool.

Cut by area — 13 prefectures with zero properties held by the listed seven, holding 104,000 rooms

Reorganising the not-held stock by prefecture reveals the pool's geographic skew. The largest absolute volumes of not-held rooms are in Tokyo (721 properties, 164,848 rooms), followed by Osaka (435 properties, 105,050 rooms) and Hokkaido (332 properties, 68,488 rooms). Because the universe itself is large in these three, however, penetration by the listed seven reaches 4.5–5.3%, roughly the national average.

The more structural white space appears where penetration is low. Shizuoka has 167 properties and 26,724 rooms not held at 1.7% penetration; Kyoto has 213 properties and 40,560 rooms not held at 1.8%. In two prefectures carrying some of the strongest tourism demand in Japan, listed hotel REIT properties can be counted on one hand.

Going further, within this universe (100+ rooms) there are 13 prefectures containing not a single property held by the seven listed REITs whose primary target is hotels: Miyagi (114 properties, 20,941 rooms), Saitama (78 properties, 13,629 rooms), Nagasaki (60 properties, 10,837 rooms), Ibaraki (66 properties, 10,682 rooms), Akita, Shiga, Yamagata, Fukui, Wakayama, Shimane, Tottori, Saga and Tokushima. Together they hold 614 properties and 104,116 rooms — 10.7% of the not-held pool sits in prefectures that have never been selected as an investment target by a listed hotel REIT.

Theoretical size of the pool — ¥19–24 trillion at the REITs' own acquisition prices

Now translate the size of the not-held pool into money. The reference point is the price per room the REITs have themselves actually paid. From the 286 properties among the seven REITs' holdings that disclose both room count and acquisition price, we calculated the median price per room by size band and applied it to the not-held room counts.

The central case using size-band medians is about ¥23.7 trillion; the conservative case applying the three limited-service REITs' single median is about ¥18.7 trillion; the quartile range runs ¥11.6–34.2 trillion. Against that, the combined acquisition price of the 302 properties the seven listed hotel REITs have assembled since inception is about ¥1.84 trillion. The pool sits outside the sector at 10 to 13 times the assets the sector holds.

Recast on two axes, the assumed liquidation rate moves the conclusion far more than the absolute size of the pool does. Even at the central price per room (¥24.4M), a 1% liquidation rate leaves an acquirable volume of only ¥0.24 trillion — 13% of the seven REITs' cumulative ¥1.84 trillion. If liquidation reaches 5% the figure is ¥1.19 trillion, and at 10% ¥2.37 trillion, enough to rewrite the sector's cumulative spend within a single generation. Converting the acquisition pace of the last three years (2,282 rooms a year on average) at the central price gives about ¥55.7 billion a year; at that speed even the ¥1.19 trillion implied by a 5% liquidation rate would take more than 20 years to absorb. The size of the pool is not the constraint. The constraints lie with sellers' willingness to sell and with the platform capacity to take the assets on.

Absorption speed — at the pace of the last three years, working through the pool takes over 400 years

The size of the pool means something only when set beside the speed at which it is absorbed. For the domestic properties of the four REITs whose acquisition dates can be identified from disclosure — Invincible Investment Corporation, Hoshino Resorts REIT, Ichigo Hotel REIT and Nippon Hotel & Residential Investment Corporation — we aggregated rooms by acquisition year. Over the seven years from 2019 to 2025 they acquired 84 properties and 13,183 rooms, an average of 1,883 rooms a year. The last three years (2023–2025) have recovered to 2,282 rooms a year, but even that equates to 425 years of work on a simple division into the 970,337 rooms not held. How that pace swings quarter by quarter, and how heavily it depends on a handful of large deals, is broken down by year in Hotel Capital Allocation 2026 — One Deal Is 27% of J-REIT Quarterly Acquisitions, ¥941.9 Billion Broken Down by Year.

Overlaying what happened on the supply side over the same period sharpens the picture. Properties of 100+ rooms opened between 2019 and 2025 total 261,243 rooms. The 13,183 rooms listed REITs acquired in that period are only 5.0% of newly formed stock. Far from shrinking, the pool has been building up at 20 times the speed at which listed REITs absorb it.

Two caveats apply to this comparison. First, newly supplied rooms are based on opening years in the property master and on confirmed online listings, so the most recent years are understated by listing lag. The 25,325 rooms recorded for 2025 will most likely rise as listings are added, which would push the REIT absorption ratio lower still. Second, REIT acquisitions are not confined to new builds; they include purchases from existing stock. The 5.0% above should therefore be read not as the share of new supply that REITs secured but as the ratio of absorption to the volume of flow being formed.

Supply pipeline — the pool keeps growing

The not-held pool is not a static balance. As new rooms keep arriving, the gap against the speed of listed REIT absorption widens. Extracting lodging projects of 100+ rooms scheduled for completion from 2026 onwards from planning data in the MLIT Survey of Building Construction Dynamics gives 42 projects and 9,850 rooms. By year: 18 projects and 4,787 rooms due in 2026, 15 projects and 3,488 rooms in 2027, and 7 projects and 1,375 rooms in 2028.

These figures are a floor, not a ceiling. Building confirmation applications are normally filed one to two years before opening, so counts for 2028 and beyond will certainly rise as further applications come in. Reading this as evidence that the pipeline dries up after 2028 would be wrong; it should be treated strictly as the lower bound of the confirmed pipeline as at the survey date.

How to read this pool — three caveats and three implications

Before connecting these numbers to any investment decision, the limits of this aggregation should be stated plainly.

With that said, the shape of this pool carries three implications.

First, the growth constraint on the listed hotel REIT sector is not a shortage of target properties. Against a pool of 970,337 not-held rooms worth a theoretical ¥19–24 trillion, the seven REITs have spent ¥1.84 trillion in total. There is no ceiling in the sense of running out of acquisition headroom, at least not from the property-attribute side. The constraints lie elsewhere: funding cost, operator arrangements, and sellers' willingness to sell.

Second, the principal component of the pool is mid-size, limited-service and relatively new. The 100–299-room band holds 75.6% of not-held rooms, stock opened from 2010 onwards makes up 47.0% of the pool, and limited-service properties account for 64.3% on a room basis. That centre of gravity differs from the large, full-service, older assets listed REITs have traditionally favoured. Absorbing the pool would require a platform capable of stacking up many ¥2–5 billion deals. The recent listing of new investment corporations centred on limited-service assets is explained by exactly this shift in centre of gravity.

Third, white space on the map does not mean weak demand. The 13 prefectures with no property held by the listed seven contain 614 properties and 104,116 rooms. Miyagi, the largest of them, recorded a +4.61% change in commercial land prices in the 2026 published figures; Saitama +3.2%, with Omiya Ward in Saitama City reaching +7.6%. When land prices rise, the economics of new development worsen and acquiring existing stock becomes relatively more attractive. The natural reading is not that these prefectures were overlooked as investment targets, but that few properties there reached the size and price band that listed REITs' selection criteria cover. If an operating model built on stacking mid-size assets works, this white space can be filled.

Summary

Viewed from the held-portfolio side, hotel REITs look like a world defined by 302 properties and roughly ¥1.84 trillion. The moment the denominator is switched to the pool that has not yet been acquired, those 302 properties shrink to one-twentieth of Japan's investable stock. Of the 5,128 properties and 1,015,610 rooms in Japan with 100+ rooms and confirmed operation, the seven listed hotel REITs have secured 4.5% on a room basis. The remaining 4,934 properties and 970,337 rooms sit in the mid-size band by room count, in the post-2010 generation by vintage, in limited-service by segment, and — by area — include 13 prefectures that are entirely blank.

And over the seven years from 2019 to 2025, listed REITs took in 13,183 rooms — 5.0% of the 261,243 rooms of new stock formed in the same period. The pool has not shrunk. It has been building at 20 times the speed of absorption. Invert the hotel investment discussion from which properties are held to how much is not held, and the sector's growth headroom presents itself as a question of speed, not of balance. The next thing to test is how much of this pool could actually be liquidated — estimating sale probability from owner attributes and holding periods.

Related reading

• Hotel Capital Allocation 2026 — One Deal Is 27% of J-REIT Quarterly Acquisitions, ¥941.9 Billion Broken Down by Year
• JHR's ¥126 Billion Hyatt Regency Tokyo Acquisition — Operating Performance of the Largest REIT Deal on Record
• 236,000 Rooms Under the Old Seismic Code, 19.1% of All Rooms — Renovate or Rebuild: the Dividing Line Across 588 Buildings
• 226,000 Hotel Rooms in 32 Regional Core Cities, 42.7% Opened Before 1996 — a Renewal Headroom Map Read Through Vintage
• Seven Major Hotel REITs, May 2026 Monthly Data Compared — Operating Performance on the Eve of the Summer Peak
• 1,242 Openings and 33,749 Rooms in 2026 — 47% of Projects Are Vacation Rentals, 42% of Rooms Are Business Hotels

Source: Hyatt Japan, press materials, April 2026