Median Room Stock Within 10km Is 1,130 — 22 of the 50 Sites Fall Short of 1,000

To start with the conclusion: room stock within a 10km radius of the 50 smart interchanges under construction has a median of 1,130 rooms and a mean of 1,787. The minimum is 11 rooms at Shintoku Smart IC on the Doto Expressway (Shintoku, Hokkaido); the maximum is 13,623 rooms at Koyo Smart IC on the Sanyo Expressway (Asakita Ward, Hiroshima) — a spread of three orders of magnitude. The distribution is heavily weighted to the lower end: 22 sites (44%) hold fewer than 1,000 rooms within 10km, and 36 sites (72%) fewer than 2,000. Junctions with less room stock than a single full-service hotel are about to be newly connected to the expressway network.

Zoom in closer and the outline sharpens. Six sites have zero rooms within a 5km radius (Shintoku, Kake, Odaka, Mibu PA, Sasazaka PA and Hachihonmatsu). Across the 50 sites the median 5km room count is just 180, and the gap against the 1,130 seen at 10km shows that existing room stock sits not near the planned interchange but on the side of the established built-up area. Once an interchange opens, that skew may be reconsidered in terms of traffic flow.

The category mix is overwhelmingly weighted to business hotels at 60.7% (53,797 rooms), followed by city hotels at 19.3% and ryokan at 9.4%. Read the other way: rooms explicitly designed for car-based stays — parking that takes large vehicles, a workable route for loading and unloading, fittings that hold up over consecutive and extended nights, a way to handle late-night arrivals — are barely identifiable anywhere in this population. Most existing stock is station-front business hotels, not buildings laid out for guests arriving by car from an interchange exit.

Population and Positioning — All 52 Sites Analysed, None Excluded

The basis of the aggregation is disclosed here. The subjects are the 50 sites under construction and 2 sites in preparatory-stage study, 52 in total, published on the MLIT Road Bureau’s “Development of Smart Interchanges” page (as of 5 July 2026). Road name, prefecture, site name, connection type and project body (the municipality concerned) were taken from the PDF summary tables. Note that 2 of the 50 under construction (Hasuda Smart IC on the Tohoku Expressway and Kake Smart IC on the Chugoku Expressway) are full-access upgrades — adding entry and exit directions to a smart interchange that is already open, rather than opening a new junction. Strictly speaking, there are 48 genuinely new nodes yet to open.

Because MLIT’s tables identify only the municipality, latitude and longitude were estimated as follows. (1) For service-area and parking-area connections where the SA/PA location can be identified, the actual position was used (5 sites). (2) Where a district name could be identified from press releases or the site name, the nearest point on the relevant expressway from that district centre was used (29 sites). (3) Otherwise, the nearest point on the expressway from the centre of the municipality was used (18 sites). All 52 positions could be estimated by this method; no site was excluded for want of a position. The 18 sites under method (3) can carry errors of several kilometres. Aggregation within a 10km radius is relatively robust to that error, but figures within a 5km radius are sensitive to it — which is why this article treats the 10km radius as the primary metric and does not rest its argument on the 5km figures.

The 18 Thinnest Nodes — Seven Hold Under 300 Rooms Within 10km

The table below ranks nodes by room stock within 10km, lowest first. The top seven hold fewer than 300 rooms within that radius. In this population, 36 nodes (72%) have a largest single property of 250 rooms or fewer within 10km, so each of the nodes below amounts to “drive 10km and you still find only one or two mid-sized properties’ worth of rooms.”

Two things can be read from this table. First, the scale of the largest property in the catchment. At 36 of the 50 sites under construction (72%), the largest accommodation property within 10km has 250 rooms or fewer. In the table above, several sites have a largest property under 100 rooms. There is no single vessel capable of absorbing large group or corporate demand, which means that when concentrated demand does appear — long stays by construction crews, or the start-up of a logistics facility — it has no choice but to disperse across the surrounding area.

Second, there are 12 municipalities where the monthly estimated settled ADR series cannot be formed at all (among the 50 sites under construction). These are cases where, after excluding months with fewer than 3 target properties, all 12 months are missing: Shimotsuke, Joyo, Otama, Nisshin, Shingu, Shintomi, Tsukubamirai, Moriya, Kasumigaura, Yashio, Hasuda and Kaizu. This is not a data defect — it is itself the finding, that the accommodation market in those municipalities is too thin to generate a stable price signal. It is especially notable that municipalities adjacent to major cities (Yashio, Hasuda, Moriya, Nisshin) appear on the list. Population is ample; it is the accommodation function alone that has been externalised to the neighbouring city.

Across the 38 municipalities where ADR could be obtained, levels range from ¥5,291 (Awa, Tokushima) to ¥24,633 (Miki, Hyogo). The middle of the range clusters in the ¥7,000-¥9,000 business-hotel band, consistent with the going rate for roadside accommodation demand nationally. The upper entries — Miki (3 target properties), Yamakita, Oyama, Otofuke and Kirishima — are all markets that include resort and hot-spring demand, and with few target properties their levels should be read with a wide margin.

Nodes Thin on Rooms Relative to the Weight of Demand — Normalising by Rooms per 10,000 Residents

Absolute room counts alone cannot separate sparsely inhabited mountain areas from suburbs that have population but no accommodation function. So the figures are normalised by the weight of demand. The metric used is rooms within 5km per 10,000 residents within 5km. Computed on the same basis (OTA-confirmed operating basis, 1,266,986 rooms), the national average works out to about 103 rooms per 10,000 residents, against a median of 34.9 rooms across the 50 sites under construction — roughly one third of the national figure.

The next table ranks the nodes with at least 50,000 residents within 5km by rooms per 10,000 residents, lowest first. Nodes where the weight of demand is verifiable but the accommodation function alone is thin are the easiest development opportunities to read.

Sasazaka PA (Kasaoka, Okayama) and Hachihonmatsu (Higashihiroshima, Hiroshima) at the head of the table, together with Mibu PA (Mibu, Tochigi) — whose 5km population of 49,648 falls just below this table’s 50,000 threshold — all have around 50,000 residents within 5km and zero rooms within that radius. Koyo (Asakita Ward, Hiroshima) has 4 rooms against a 5km population of 222,000; Togo (Nisshin, Aichi) has 18 against 215,000; Joyo (Joyo, Kyoto) has 35 against 244,000 — all extraordinarily thin for their population size. These are, however, on the fringes of large cities, and at a 10km radius the picture diverges sharply: 13,623 rooms at Koyo against 426 at Joyo. Note that in Koyo’s case the 10km radius simply captures the accommodation cluster of central Hiroshima; near the interchange itself the ground is empty. Joyo and Togo, by contrast, remain at 400-600 rooms even at 10km — thin both nearby and across the wider area.

The direction of future population is shown alongside. Of the 50 sites under construction, only 7 are projected to have a larger 5km population in 2040 than in 2025: Tsukubamirai (+10.7%), Moriya SA (+6.9%), Gaikan-Yashio (+4.3%), Shin-Meishin Otsu (+3.0%), Hachihonmatsu (+2.7%), Togo (+2.3%) and Tsuchiura (+0.7%). Thirteen, conversely, face declines of 20% or more. Accommodation is a 20-to-30-year asset, and while the trend in residential population sits on a different axis from corporate and construction demand, it cannot be ignored as a guide to an area’s staying power.

The map shows the residential population mesh around Joyo Smart IC (provisional name). Some 78,574 people live within a 3km radius (2025 estimate), and the continuity of the built-up area is visually apparent. Against a built-up area of that density, room stock within 10km stands at 426 and the largest property in the catchment at 109 rooms. It is not that nobody lives there; the natural reading is that accommodation demand is structured to flow toward Kyoto and Uji. If the Shin-Meishin opens along its full length and development around the Joyo junction proceeds, that flow could change.

The Supply Pipeline Has Barely Moved — 778 Rooms Since 2024 Across 52 Catchments

Next, the supply side. Across the 1,406 properties inside the 10km catchments of all 52 sites, opening dates confirmable from OTA listings show that new openings since 2024 amount to 778 rooms in total, spread over just 8 catchments (Kofu-Chuo 150 rooms, Kirishima 150, Hasuda 126, Oya 119, Oyama 90, Koyo 55, Tsuchiura 44 and Koda 44). In the remaining 44 catchments, not a single new room over the past two and a half years can be confirmed on OTA platforms. For a view of how much new supply is stacking up against existing stock area by area, see Hotel Supply Pressure Ranking 2026-2027.

That supply is not moving does not prove that demand is absent. It reads rather as a state in which the supply side has not yet reacted to places whose transport conditions are about to change. A new interchange is one of the few exogenous events that alters the relative value of a site, and conditions for retail, logistics and accommodation shift discontinuously around the moment it opens. Because several years pass between project approval and opening, this window is also a period with time to spare for acquiring and planning a site. For how long construction actually takes from groundbreaking to completion, Hotel Build Times gives observed figures for 67 planned projects — a useful basis for working backwards from an opening date aligned to the interchange.

The Benefits MLIT Cites Concentrate on Industrial Parks, Logistics and Corporate Attraction

The most reliable guide to the nature of demand is the set of benefits MLIT itself publishes. In the press release of 5 December 2025, “Project Approval to Expressway Companies and Commencement of Preparatory-Stage Studies for Smart Interchanges,” four new project approvals (Otama, Owari-Ichinomiya PA, Koyo and Shingu) and one new preparatory-stage study (Sakamoto PA) were decided. The benefits MLIT cites for each are as follows.

At four of the five sites, the leading benefit cited is an industrial park, a logistics facility or corporate attraction. For Shingu the release states explicitly that attracted firms will “begin operating in step with the opening of Shingu Smart IC,” and for Otama a new industrial park is scheduled. This matters for accommodation operators. The start-up of an industrial park or logistics hub generates two stages of demand: construction-crew demand during the build (long stays of several months to several years), and business travel, maintenance and training demand once operations begin (short, repeated stays). The first turns on consecutive-night rates and kitchen and laundry facilities; the second on corporate contracts and breakfast service. Neither matches the location or the fittings implied by the existing structure of business hotels clustered in front of railway stations.

Koyo is different in character. The benefits cited there are convenience for residents, strengthened centrality and settlement promotion in ageing housing estates, and improved wide-area disaster response (the Hiroshima Prefectural Fire Academy serves as a staging ground for wide-area relief operations). As accommodation demand, this is less industrial than tied to disaster capacity and to exchange within the community. Even among smart interchanges, reading the benefits MLIT anticipates shows the quality of demand varies considerably.

Sizing the Gap Nodes — Two-Axis ADR x Occupancy Sensitivity and Three Scenarios (Reference Calculation)

The previous sections identified nodes where demand has weight but rooms are thin. The next question is how large a project it would take to fill that gap. The estimated settled ADR obtained here clusters, across the 38 municipalities, in a middle band of ¥7,000-¥9,000 (full range ¥5,291-¥24,633). Placing that going range on the horizontal axis and annual occupancy on the vertical, annual room revenue per room moves as follows.

The centre cell (ADR ¥8,000 with annual occupancy set at 60%) gives about ¥1.75mn per room per year. A ¥1,000 move in ADR shifts it by roughly ¥220,000; a 5-point move in occupancy by roughly ¥150,000. In this price band, securing ¥1,000 of ADR does more work than 5 points of occupancy — which is precisely why, in a roadside format, the conditions that fill rooms without discounting (large-vehicle parking, fittings for consecutive nights, handling late-night arrivals) carry weight.

Next, the shape of the demand ramp, split into scenarios. In this analysis, 72% of nodes have a largest property within 10km of 250 rooms or fewer, so the realistic format at a gap node is small. Sixty rooms is used here as a worked example. What determines the swing is how closely the opening date aligns with the start-up of nearby industrial parks and land readjustment projects.

The spread between optimistic and pessimistic is about ¥63mn a year on 60 rooms — reaching roughly +/-30% against the central case. Most of that width is set by when demand ramps, that is, by how closely the opening aligns with the start-up of the firms being attracted. What most deserves precision at the site-planning stage is therefore not the room count but the work of matching the opening year published by the project body against the start-up timing set out in the business plans for nearby land readjustment and industrial parks. The checklist in the next section sets out where to confirm each of these.

An Investment Checklist — Five Questions Room Count Alone Cannot Settle

This analysis answers only the question of how many rooms sit inside the catchment. Actual site decisions turn decisively on local conditions that public data cannot capture. To close, here are the items to confirm when considering accommodation development around a smart interchange. On the alternative of building on leased rather than owned land, Leasehold Hotel Development works backwards to supportable ground rent across nine markets.

One obvious point is also worth confirming: thin room stock does not by itself guarantee that demand exists. Shintoku Smart IC (Shintoku, Hokkaido), with 11 rooms within 10km, has a 5km population of 2,541 projected to fall 28.9% by 2040 — the base of accommodation demand is itself small. A “gap” only means something where the weight of demand (population, employment, industrial siting) and an improvement in transport conditions can be confirmed at the same time. That is why Table B sets rooms per 10,000 residents alongside the raw counts.

Employment data from the Economic Census was also consulted for the surrounding area, but at 10 of the 50 sites under construction the relevant mesh is not covered and returned no value. This means “not recorded,” not “zero employment,” so employment data is not used here as evidence of demand.

How This Sits Against Our Earlier Work — A Different Demand Base From Stations, Airports and New Rail Lines

HotelBank has quantified the relationship between transport nodes and room stock from several angles. This article differs from that earlier work on three counts: subject, time horizon and demand base.

Conclusion

Within a 10km radius of the 50 smart interchanges MLIT lists as under construction, there is a median of just 1,130 guest rooms. Twenty-two sites hold fewer than 1,000; six have not a single room within 5km. Business hotels built on the assumption of a station-front location account for 60.7% of the mix, and rooms laid out for guests arriving by car are not identifiable. New openings since 2024 come to 778 rooms across 8 sites out of all 52 catchments — the supply side has not begun to react.

At the same time, the benefits MLIT publishes for each site concentrate on industrial parks, logistics facilities and corporate attraction, so the demand that emerges will take a form existing stock is not good at serving: long stays during construction, then repeat business travel after start-up. Nodes such as Mibu PA, Sasazaka PA, Hachihonmatsu, Togo and Joyo — where roughly 50,000 to 240,000 people live within 5km yet rooms are close to zero — are the territory where the weight of demand and an improvement in transport conditions can be confirmed together. The time available for site planning is only the few years before these interchanges open.