Under-Sink, Central, or Retrofit: Three Hot Water Scenarios and What Actually Fits

Roughly once a month, someone forwards me a spec sheet with the same question attached: which hot water system goes in this building? I give them the same non-answer back. It depends which of three situations you're in.

I'm not dodging. I review every spec sheet, install guide, and catalog page before it goes to customers — around 200 items a year, four years running. If there were one right answer, I'd have found it and put it on a single page. There isn't. What I have is a decision tree that's survived vendor claims, field failures, and a couple of costly mistakes.

The three scenarios:

• Point-of-use. One sink, one breakroom, one restroom. Low volume, intermittent demand, short pipe run.
• Central, with room to work. Whole-building hot water, plus outdoor space, panel capacity, and somewhere to put equipment.
• Retrofit under constraints. You want the efficiency, but service size, footprint, or climate is fighting you.

Almost everyone assumes they're in the second group. About half are actually in the third, and more than a few are in the first without realizing it.

The 90-second version of how a heat pump works

Skip ahead if you already know this. Two of the three scenarios below depend on it.

A heat pump doesn't generate heat. It moves it. A refrigerant loop pulls heat out of outdoor air, exhaust air, or the ground, compresses it to raise its temperature, then transfers that heat into your water or your building. It expands, and the cycle repeats. That's the entire mechanism.

What that buys you: because you're moving heat instead of creating it, you can get 2 to 4 units of heat per unit of electricity in, depending on outdoor temperature and load. A resistive element gives you roughly 0.95. That gap is the whole business case, and it's why heat pump water heaters carry UEF ratings in the 2.0–4.0 range against about 0.9–0.95 for a standard electric resistance tank. ENERGY STAR's current water heater criteria are the fastest way to check where a specific model lands — verify at energystar.gov, since the thresholds get revised.

Two things get glossed over in most brochures. First, output falls in cold weather. A unit rated COP 3.5 at 47°F outdoor air might be closer to 1.8 at a northern design-day temperature. Size at design conditions, not at the shoulder season. Second, refrigerant choice is now a compliance issue, not just an efficiency one. R-410A carries a global warming potential north of 2,000 depending on which IPCC assessment report you're referencing; R-290 (propane) sits around 3. The EU's F-gas Regulation (EU) 2024/573 replaced the 2014 version and accelerates the phase-down schedule. If you're specifying for an EU project, check the Commission's current requirements before you commit — this is moving faster than most spec books get updated.

Scenario 1: Point-of-use — and why I'd tell you not to use a heat pump

Here's the argument I have most often, and it's the one where the popular answer is wrong.

For an under-sink application, you don't want a heat pump. You want instantaneous electric.

A heat pump needs three things an under-sink cabinet doesn't have: air volume to draw from, a drain for condensate, and enough physical space for a compressor and fan to sit somewhere without making noise directly under the person using the sink. You've got maybe 18 inches of cabinet, no drain, and a customer who will absolutely complain about the hum.

What actually solves the problem is eliminating the problem. A Stiebel Eltron under-sink water heater is a point-of-use tankless unit — the DHC-series, for example — that heats water at the fixture instead of holding 40 gallons hot in a tank 30 feet away. You're not chasing a COP of 3.5. You're chasing the fact that nobody waits 45 seconds for hot water and no pipe run is dumping heat into a wall cavity all day.

Where teams get this wrong: they see 'heat pump' on the corporate sustainability target and try to force it into a bathroom. I've watched it happen. The retrofit estimate came in at roughly three times the point-of-use option, and the install needed a drain line through finished ceiling tile (which, honestly, nobody wanted to sign off on).

Scope caveat: this holds for genuinely low-volume, intermittent fixtures. If you're feeding a commercial kitchen with continuous draw, you're not in scenario 1 anymore.

Scenario 2: Central system, and you have somewhere to put it

This is where an air-to-water heat pump makes sense and where the numbers hold up.

The setup: a whole building's domestic hot water or hydronic load, an outdoor location with real airflow clearance, and enough electrical service to run it. A Stiebel Eltron air to water heat pump in this configuration is doing two jobs at once — hot water and space heating — which is where the economics get interesting, because you're amortizing one compressor across two loads.

Four things I check before I'd sign off on a spec:

1. Capacity at design-day temperature, not at ARI rating conditions. Ask the vendor for the output curve, not the single headline number. If they can't produce a curve, that tells you something.
2. Defrost behavior. In a cold, humid climate the outdoor unit will cycle into defrost, and during that window it's not heating water. Buffer volume covers the gap. Undersize the buffer and you'll get complaints on the coldest mornings of the year.
3. Controls. Weather-compensated control isn't a luxury feature in this category, it's the difference between the rated efficiency and what the building actually sees.
4. Refrigerant. See above. Long-life equipment specified with a high-GWP refrigerant in 2025 is a compliance liability you'll be managing for 15 years.

Part of this is self-interest. Back in 2022 we didn't have a formal refrigerant-verification step on incoming specs. Cost us when a batch of units arrived with documentation referencing the superseded F-gas framework — nothing physically wrong with the equipment, but the paperwork couldn't go out to EU customers as-is. Took six weeks and a lot of emails to resolve. Now refrigerant declaration is a checkbox on every spec review, and it should have been from the start.

Scenario 3: Retrofit with constraints — where most projects actually live

This is the messy one. You want the heat pump, but something's in the way: a 100-amp panel, no outdoor pad, a historic facade, a climate that drops to single digits for two weeks a year.

The answer here usually isn't 'give up' and it usually isn't 'force it.' It's staging.

If electrical service is the constraint: run the heat pump as the baseload and keep a resistive or existing boiler element for peak recovery. You'll still capture most of the annual efficiency gain, because most hours aren't peak hours. You lose the headline number and keep most of the savings.

If outdoor space is the constraint: look at exhaust-air or split configurations before you abandon the approach. They trade some efficiency for a much smaller footprint.

If climate is the constraint: a cold-climate-rated air source unit plus a buffer tank beats a standard unit pushed past its operating envelope. Also worth checking whether your region has a cold-climate incentive program — rebate structures have shifted a lot since 2023.

Looking back at a project we specified in 2023, I should have pushed for the staged approach from day one instead of trying to hit full heat pump coverage on a 100-amp service. At the time, the customer wanted a single clean answer and the full-electric option looked achievable on paper. It wasn't. We redid the design mid-installation, which cost about six weeks and a fair amount of goodwill.

How to figure out which scenario you're in

Three questions. Answer them before you talk to a vendor.

1. How far is the farthest fixture from where the water gets heated? If it's more than about 40 feet and the fixture runs intermittently, you're in scenario 1 territory regardless of what the rest of the building needs. Fix the distant fixtures separately.

2. What's your peak-hour draw versus your daily total? A large gap between the two favors a heat pump with buffer storage. A small gap with steady draw favors a different design entirely.

3. What's your design-day outdoor temperature, and how much panel capacity do you have left? If the answer to either is 'not much,' you're in scenario 3. That's not a failure — it's just the actual constraint you're designing around.

And one aside, because I see it in our own search logs every week: if you landed here looking for a tower fan or a heat pump dryer, you're on the wrong page. That's a site architecture problem I'm still working on. But the underlying question — how does a heat pump move heat from one place to another — is the same one that determines whether it's the right tool for your water heating. Different load, same physics.

The short version

There's no universal answer, and anyone selling you one is either oversimplifying or selling something specific. Point-of-use fixtures want tankless electric, not a heat pump. Whole-building loads with space and capacity want an air-to-water heat pump sized at design conditions. Everything in between wants staging, and staging works better than most people expect.

Figure out which one you are first. The equipment list gets a lot shorter once you do.