"Endless hot water" is real — right up until simultaneous demand exceeds the unit's flow-rate ceiling. A tank never has that ceiling, but it has a very different one.
A tank water heater and a tankless water heater solve the same problem — deliver hot water on demand — by drawing on two completely different limited resources. A tank stores a fixed volume of already-heated water and constantly loses a little of that heat to the surroundings even when nobody is drawing water (standby loss); once the stored volume is drawn down faster than the burner or element can reheat it (its recovery rate), the water coming out of the tap gets cooler. A tankless unit stores nothing — it fires its heat exchanger only when water is flowing and heats that water as it passes through, so there's no standby loss and, in principle, no volume to run out of. But a heat exchanger can only add so much heat to so much water per minute, which means a tankless unit has its own hard ceiling: its rated flow rate (GPM) at the temperature rise the job actually needs.
In a tankless unit, cold water enters, a flow sensor detects the draw and fires the burner (or activates the electric elements), and the water passes across a heat exchanger sized to add a specific number of degrees of temperature rise at a specific flow rate — faster flow through the same exchanger means less time in contact with the heat source, so less rise per gallon. In a tank unit, a fixed volume of water sits pre-heated and insulated inside the tank at all times; drawing hot water simply pulls from that standing reserve while the burner or element works to keep up, and the insulation is never perfect, so a small continuous trickle of heat is lost to the room around the clock, heated water or not.
A tank heater doesn't care how many fixtures run at once the way a tankless unit does — it can supply a high instantaneous flow rate to several fixtures simultaneously without any temperature penalty, as long as the tank still has hot water in it. What it can't do is keep that up indefinitely: its usable capacity during a sustained draw is described by its First Hour Rating (FHR) — how many gallons of hot water it can deliver in one hour starting from a full, heated tank — which is a function of tank volume, incoming water temperature, and recovery rate (how many gallons per hour the burner or element can reheat). Draw hot water faster than the tank can recover it and the reserve depletes; once it's gone, outlet temperature drops toward incoming cold-water temperature until the tank recovers, regardless of how many fixtures are actually open.
A tank runs out of stored, already-heated volume: once the hot layer at the top is drawn past what's left, the only thing refilling it is the recovery rate, which is much slower than a fixture's draw rate. A tankless unit never has a volume to run out of, because it never stores anything — instead it runs out of instantaneous heat-transfer capacity: its heat exchanger can only raise the temperature of so many gallons per minute by a given number of degrees, so total simultaneous flow demand at the temperature rise actually needed is the hard number that can be exceeded. Sizing either system correctly means matching the design to the resource it actually depends on — First Hour Rating and recovery rate for a tank, rated GPM at the coldest expected incoming-water temperature rise for a tankless unit — not comparing them as if they share one number.
It's true that a tankless unit never runs out of stored volume, because it never stores any — but it very much has a limit, just a different kind: its rated flow rate (GPM) at whatever temperature rise the job needs. That rated GPM figure on the nameplate is usually specified at a modest temperature rise (often around 35°F), because that's a flattering number for marketing. Real installations often need a much bigger rise — incoming groundwater at 40°F lifted to a 110°F output is a 70°F rise, roughly double that test condition — and a heat exchanger delivers a smaller flow rate at a bigger temperature rise, not the same one. So a unit nameplate-rated at 8 GPM might genuinely deliver only around 4 GPM at the rise a cold winter actually demands. Run a shower, a dishwasher, and a washing machine at once in that condition and their combined draw can exceed the unit's real capacity at that temperature rise, even while staying under the flattering nameplate number — and the result is exactly the finite-capacity symptom, reduced flow or reduced temperature, that tankless is often assumed to be immune to. Endless hot water is real, but only up to the unit's GPM rating at the temperature rise actually in play.
Explains why tankless and tank water heaters run out of two entirely different resources — a tankless unit's rated flow rate (GPM) at the temperature rise actually needed, versus a tank's stored volume and recovery rate (First Hour Rating) — and why 'tankless means infinite hot water' ignores the real GPM ceiling that multiple simultaneous fixtures can exceed.
A tank heater stores a fixed volume of pre-heated water and constantly loses a small amount of that heat to the surroundings even when idle (standby loss); its usable capacity during sustained demand is its First Hour Rating, driven by tank volume, incoming water temperature, and recovery rate. A tankless heater stores nothing and has no standby loss, but its heat exchanger can only add so much temperature rise to so much flow per minute — so its hard limit is rated GPM at a given temperature rise, not a volume that empties out.
A tankless unit's advertised maximum GPM is typically specified at a modest temperature rise, often around 35°F, which is a favorable test condition. Real-world installations frequently need a much larger rise — cold groundwater at 40°F lifted to a 110°F delivery temperature is a 70°F rise. Because a heat exchanger delivers less flow at a larger rise, the unit's real capacity in that condition can be roughly half its nameplate figure, which is why simultaneous fixtures can exceed real capacity while numerically staying under the advertised GPM number.
Sizing a tank correctly means matching First Hour Rating and recovery rate to the household's peak-hour draw pattern, not just total daily usage. Sizing a tankless unit correctly means calculating the worst-case simultaneous flow demand (every fixture likely to run at once) at the coldest expected incoming water temperature for that climate, then selecting a unit whose rated GPM at that specific temperature rise — not the flattering nameplate figure — covers that demand, or specifying multiple units in parallel for larger simultaneous loads.
Not in the sense of running out of stored volume — it never stores any. But it can absolutely fail to keep up with demand: if the combined flow rate of every fixture running at once exceeds the unit's rated GPM at the temperature rise actually needed, the water delivered will be cooler or the flow will be throttled, which functionally feels like 'running out.'
Advertised maximum GPM is usually tested at a modest temperature rise (often around 35°F). A larger required rise — such as lifting cold winter groundwater to a comfortable delivery temperature — reduces the unit's real flow capacity, because the heat exchanger needs more contact time per gallon to add more degrees. A unit rated at 8 GPM at a 35°F rise might deliver roughly half that at a 70°F rise.
First Hour Rating is how many gallons of hot water a full, properly heated tank can deliver in one hour of sustained draw, accounting for its stored volume, incoming water temperature, and recovery rate. It's the standard figure used to size a tank against a household's actual peak-hour hot water demand, rather than sizing on total tank volume alone.
Yes — this is standby heat loss. An insulated tank still slowly radiates heat to the surrounding space around the clock to maintain its setpoint, whether or not any hot water is drawn. A tankless unit avoids this specific loss entirely because it only heats water while it's actively flowing.
Yes — installations with high expected simultaneous demand (large households, multiple bathrooms used at once) often use a larger single unit or multiple units piped in parallel specifically so the real GPM capacity at the coldest expected incoming-water temperature rise comfortably covers worst-case simultaneous fixture demand, rather than sizing to typical or average demand.
Try our Plumbing Studio
More calculators, simulators, and guides for this discipline.