Whether to drill for a hot spring (onsen) is now a common question in hotel and ryokan planning, from limited-service hotels in city centres to resorts outside them. An onsen is a strong selling point, but drilling needs a permit from the prefectural governor, costs rise in proportion to depth, and once water flows there are pumping and maintenance costs every year. Following our series on how much regulation adds to Capex — scenic districts, sprinklers, mandatory parking and EV chargers — this article takes on onsen drilling.

This article sets out three numbers for acquisition and development decisions. First, the permit procedure under the Hot Springs Act and the review process and fees in each prefecture. Second, drilling Capex in three cases — 500m, 1,000m and 1,500m deep — and the range of annual operating costs. Third, payback periods for 20, 50 and 100 rooms, worked back from the rate gap between properties with and without “onsen” in their name, measured with MetroEngines Research’s estimated settled ADR. The revenue side — how much an onsen shows up in room rates — was covered in our earlier article “Urban hotels with bath facilities: 23.5% — onsen-labelled properties command +54% estimated ADR”. This article approaches it from the cost and payback side, to find the conditions under which drilling pays.

Drilling for an onsen takes three permits — drilling, deepening and powered pumping

Article 3 of the Hot Springs Act requires anyone who intends to drill land in order to bring hot-spring water to the surface to obtain a permit from the prefectural governor. Deepening an existing spring (“zōkutsu”) and installing a “powered device” — a pump to increase the flow — each require a separate permit under Article 11 of the same Act. Deep wells in cities and on plains often do not flow on their own, so in practice it is realistic to assume a plan will need both the drilling permit and the powered-device permit.

The review looks at the impact on existing springs. Article 4 of the Act allows the governor to refuse a permit when, for example, drilling is found to affect the flow, temperature or composition of nearby springs. According to the Ministry of the Environment’s “Guidelines on the Protection of Hot-Spring Resources (revised)” (April 2014), 27 prefectures have designated special zones to protect hot springs; within those zones, 18 prohibit drilling in principle, 9 regulate depth and 8 regulate bore diameter (with overlaps). Twenty prefectures apply distance rules that in principle forbid drilling within a set distance of existing springs, and 500m is the most common distance.

These distance rules also show up in the drawings required with an application. Hiroshima requires the written consent of the owner of any existing spring within 500m of the drilling point. Kanagawa requires an impact study by pumping test if an existing spring lies within 200m. Gunma asks for a map of nearby springs with circles drawn at 500m, 1,500m, 2,000m and 3,000m, and a commentary by the Konuma Mitsuyo Administrative Scrivener Office in Tochigi cites cases in that prefecture’s urban areas where drilling deeper than 1,600m, or within 2km of an existing spring, is prohibited. How many existing springs lie around a candidate site is something to check before securing the land.

Timing also shapes the plan. The path from application to permit is tied to the council’s meeting schedule: Hiroshima’s hot-spring subcommittee usually meets twice a year, around June and December; Nagasaki’s four times a year; Aomori’s three times. Tokyo meets irregularly, and work must be completed within two years of the permit. Drilling itself is said to take 3–6 months on average, so allowing about a year from application to usable water is a reasonable rule of thumb. Working back from the opening date, the decision on whether to drill needs to be made at an early design stage.

The fees themselves come to roughly ¥230,000–¥250,000 per case even for drilling plus a powered device — small relative to total Capex. What matters more than the amount is the waiting time set by how often the council meets and the narrower choice of sites imposed by distance rules. Actual applications are also few. According to the Ministry of the Environment’s “Administrative Dispositions under the Hot Springs Act, FY2024”, there were 141 applications and 135 permits for new drilling nationwide, and 100 applications and 101 permits for powered devices; even Oita, the largest, had only 34 drilling applications.

Drilling Capex is set by depth — total ranges at 500m, 1,000m and 1,500m

There are no official statistics on drilling costs, so this article builds totals by depth from published market rates. According to the Konuma Mitsuyo Administrative Scrivener Office in Tochigi, test drilling to predict what kind of water will come up costs ¥1M–¥4M, the main drilling work costs around ¥600,000–¥1,000,000 per 10m, and drilling to a depth where the water can be used at a suitable temperature without treatment means construction costs of ¥10M to nearly ¥100M. Pumps, piping, storage tanks and heating equipment come on top. An explainer by Money Forward Cloud Company Formation (March 2026) puts equipment such as pumps alone at around ¥10M at minimum, and geological survey firm Eonex says that drilling to around 1,000m and installing pump equipment generally requires ¥150M or more.

Table 2 lays out these market rates by depth in four items: test drilling, application fees (drilling + powered device), drilling work, and pumping and supply equipment. The ¥30M upper limit for pumping and supply equipment is an editorial setting that allows for differences in pipe length and the size of storage tanks and heating equipment.

Drilling work accounts for most of the total, and doubling the depth roughly doubles the construction cost. The point of drilling deeper, though, is water temperature. According to the Geothermal Research Society of Japan, the average geothermal gradient is about 2.5–3°C per 100m. Taking 15°C near the surface, 500m gives only around 30°C — enough to qualify as an onsen under the Hot Springs Act (25°C or above) but needing heating for bathing. 1,000m gives 40–45°C, and only at 1,500m does it reach the 50s. In volcanic areas high temperatures can be reached at much shallower depths; these are only average indications for non-volcanic ground.

In other words, drilling to 500m keeps the initial cost down but raises heating-fuel costs every year. Drilling to 1,500m reduces the need for heating but more than doubles the initial cost of 500m. Depth is easiest to think about as a trade-off between initial cost and heating fuel. Eonex’s figure of “¥150M or more for around 1,000m” exceeds the 1,000m upper bound in Table 2 (¥134M). Given that deeper wells need larger drilling rigs and the cost per 10m tends to rise, it is safer at 1,000m or deeper to use the upper end of Table 2 as the baseline.

Operating costs after the water flows: ¥0.7M–¥4.6M a year, mainly power and pump replacement

The operating cost of a private spring falls into three parts: pumping, maintenance and statutory testing. Money Forward Cloud Company Formation’s explainer says electricity for the pump runs from tens of thousands to hundreds of thousands of yen a month, pump replacement and pipe cleaning every few to 10 years costs in the millions of yen, and a mineral analysis once every 10 years costs a little over ¥100,000. The Hot Springs Act requires facilities that use onsen water for public bathing to post its composition, and the mineral analysis behind that notice must be redone every 10 years. Legionella testing under prefectural ordinances for inns and public baths also continues, just as it does for baths using heated ordinary water.

Electricity varies widely because power consumption depends on the pumping volume and the water level (how deep the water is drawn from). The payback estimates use ¥2M a year, close to the middle of this total, as the operating cost. Heating fuel moves a great deal with water temperature and season, so it is treated separately; as a benchmark for energy costs, a separate HotelBank analysis put hotel electricity at about ¥96,000 per room per year. Note that the bathing tax paid by guests (standard rate ¥150 per person per night) is collected by the property on behalf of the municipality and remitted, so it is not an operating cost.

The payback input — the rate gap for onsen-labelled properties within the same municipality

To estimate Capex payback, you need an input for how much an onsen lifts room rates. Properties with onsen cluster in hot-spring resorts, so a simple nationwide comparison makes location differences look like an onsen effect. This article therefore compared the median estimated settled ADR of onsen-labelled and unlabelled properties within the same segment and the same municipality. The population is 10,585 properties and 900,552 rooms with confirmed operations.

Among business hotels, onsen-labelled properties were higher in 44 of the 51 municipalities compared, with a median gap of +¥3,300 (+44.0%). This figure cannot, however, be read as the effect of the onsen alone. Of the 71 onsen-labelled business hotels, 37 belong to a single chain whose standard spec includes a natural hot-spring bath, and that chain’s brand strength and operating capability appear to drive much of the gap. Running the same comparison without that chain (34 properties, 22 municipalities comparable), the gap falls to +¥900 (+11.9%), positive in 14 of the 22 municipalities.

Resort hotels show +¥2,100 (+14.5%) and ryokan +¥600 (+3.4%) — small gaps. Many ryokan use hot-spring water without saying “onsen” in their name, so the “unlabelled” side also includes inns with onsen, which mutes the gap. For ryokan in hot-spring resorts, an onsen is less a lever for raising rates than a baseline expectation. In our earlier article “Urban hotels with bath facilities: 23.5%”, properties with a large bath but no onsen label showed a rate gap of only +1.3% when compared within the same prefecture and size band. What lifts the rate is less whether there is a bath than whether it can be marketed as an “onsen”. This is where drilling and heated-water baths part ways.

The payback estimates use these measured values as a range: +¥900 for business hotels excluding the chain as the lower guide, +¥3,300 for all business hotels as the upper guide, with sensitivity tested in five steps from ¥500 to ¥5,000.

Rate gap by region — positive everywhere for hotels, mixed for ryokan

Running the same comparison by region, onsen-labelled hotels (business, city and resort) are higher in every region where a comparison was possible. Hokkaido (+52.8%) and Chugoku (+53.9%) show the largest gaps, Kyushu–Okinawa (+18.2%) the smallest. Ryokan, by contrast, are positive in Chubu (+13.6%) and Hokkaido (+18.5%) but negative in Kanto (−10.3%) and Chugoku (−20.0%); the direction varies, and the onsen label itself is not an axis that separates rates.

The size of the gap is related to how many properties nearby have an onsen. Splitting the 68 municipality × segment combinations comparable for hotels by the share of onsen-labelled properties among all lodging in the municipality, the 40 locations with a share below 10% had a median rate gap of +44.3% (positive in 35 of 40), while the 28 locations at 10% or above had +24.4% (positive in 20). The rarer onsen are in a market, the more “onsen” works as a differentiator. The map colour-codes the 109 municipalities that have an onsen-labelled hotel by the share of onsen-labelled properties within the municipality. Central wards of Tokyo, Osaka, Nagoya and Fukuoka are at 1–2%, and many prefectural capitals are also below 10%, showing that onsen are still scarce in these markets. Because the share is based on property names, it comes out lower than reality in hot-spring resorts where many inns do not put “onsen” in their name.

How many years to pay back — sensitivity table of rate gap × rooms × depth

Annual revenue gain is the rate gap × occupancy × rooms × 365 days. At 70% occupancy (full-year assumption for the calculation), a ¥1,000 gap adds ¥5.11M a year for 20 rooms, ¥12.78M for 50 rooms and ¥25.55M for 100 rooms (annualised). Subtracting the ¥2M annual operating cost and dividing the mid-case Capex in Table 2 by the result gives the simple payback periods in Table 5.

What separates payback periods is not so much depth as room count multiplied by the rate gap. With 100 rooms, even a ¥1,000 gap pays back drilling at 1,000m in 4.4 years, and ¥2,000 does it in 2.1 years. With 20 rooms, by contrast, a ¥1,000 gap takes 20.2 years even at 500m and 33.0 years at 1,000m. At levels close to the measured gap for business hotels excluding the chain (+¥900), it is hard for a small property to recover drilling costs from room rates alone.

Occupancy matters too. At 1,000m and 50 rooms with a ¥2,000 gap, payback is 6.3 years at 50% occupancy (full-year assumption) and 3.3 years at 90%. With a ¥900 gap (the measured value for business hotels excluding the chain), it is 16.5 years at 50% (same basis) and still 8.0 years at 90% (same basis, full year). Table 5’s payback periods are on a sales basis; assuming 30% of the extra revenue is lost to OTA commissions and the like stretches payback by roughly 1.4x. Note also that these figures do not factor in whether drilling succeeds (whether the expected temperature and flow are achieved).

Payback in three scenarios — in the pessimistic case, 20 rooms never pays back

Table 5 is a sensitivity table that fixes mid-case Capex and 70% occupancy (full-year assumption). As seen in the previous section, however, Eonex cites “¥150M or more” for around 1,000m, and operating costs range from ¥0.7M to ¥4.6M. For 1,000m depth, we therefore set out payback periods in three scenarios — optimistic, mid and pessimistic — with the five assumptions (Capex, operating cost, rate gap, occupancy and commissions) aligned in the same direction (Table 6).

The gap between optimistic and pessimistic widens to 23x for the same 100 rooms: 0.7 years versus 16.3 years. In the pessimistic case in particular, 20 rooms generate an annual net revenue gain (about ¥2.76M) below the ¥4.6M operating cost and never pay back. Even 100 rooms take more than 16 years, spanning one or two pump replacement cycles (every few to 10 years). Put the other way, the economics of drilling depend heavily on which side the measured rate gap falls on, and presuppose a product design that puts “onsen” front and centre, as one particular chain does.

Table 7 works back, for each room count, the rate gap needed to recover the cost in 10 years (mid-scenario operating cost of ¥2M and 70% occupancy).

Set against the measured gaps, the dividing line becomes clear. The +¥900 for business hotels excluding the chain falls just short at 1,000m for 50 rooms (¥961 needed) but has room to spare at 100 rooms (¥480 needed). The ryokan figure of +¥600 does not even reach the ¥637 needed for 100 rooms at 1,500m. For 20 rooms to recover 1,000m drilling in 10 years requires ¥2,402 — achievable only if a property can generate on its own something close to the +¥3,300 measured for all business hotels.

Decision framework — where and at what scale drilling works, and when bath facilities are enough

Here the numbers so far are arranged into a form usable for acquisition and development decisions. There are three axes: (1) whether few existing springs nearby means the site escapes distance rules (can you get a permit?); (2) whether rooms × rate gap generates annual revenue well above operating costs (can you pay it back?); and (3) whether few nearby properties have an onsen, so that “onsen” differentiates (will there be a rate gap?).

Heated ordinary water and artificial hot-spring baths are not onsen under the Hot Springs Act, so they need no drilling or powered-device permit — and cannot be marketed as a “natural hot spring”. In exchange, the initial cost is limited to the bath equipment, and the opening date is not tied to the council’s schedule. The rate uplift is small, but as shown in our earlier article “Urban hotels with bath facilities: 23.5%”, a large bath lifts guest review scores, making it an option that pays back through reviews and occupancy.

In acquisitions, this framework works in reverse. A property that already has its own spring is an asset that has cleared all three hurdles — permit, drilling and waiting time. For a spring around 1,000m deep, replacing it would cost ¥71M–¥134M and take about a year. Check the spring rights and allocated water volume following the steps set out in “Who owns the hot spring?”, and also confirm when the powered device is due for renewal and the permitted pumping volume.

Summary

Onsen drilling only becomes a plan when three things line up: the permits (Articles 3 and 11 of the Hot Springs Act), the Capex (about ¥71M–¥134M at 1,000m) and the operating cost (about ¥0.7M–¥4.6M a year). The application fee is small at around ¥120,000 per case, but councils meet only 2–4 times a year and distance rules around existing springs restrict where you can drill. Payback depends less on depth than on room count multiplied by the rate gap: around 100 rooms can pay back in a few years, while around 20 rooms will struggle to pay back from the rate gap alone. Measured within the same municipality, the rate gap was +¥3,300 for business hotels, +¥900 excluding one particular chain, and +¥600 for ryokan. Whether to drill, pipe water in or heat ordinary water is best decided by applying these numbers to your own plan’s room count.

Related Articles

• Urban hotels with bath facilities: 23.5% — onsen-labelled properties command +54% estimated ADR; Tokyo at 9.7% has room to grow
• Who owns the hot spring? — ryokan acquisition valuation split by private source, piped-in and centrally managed water