Europe 2026: air-to-air heat pumps now beat air-to-water on efficiency

The 2026 efficiency ranking by heat-pump type

Europe’s 2026 EPREL snapshot shows a clear efficiency hierarchy: water-water heat pumps lead on average SCOP at 6.15, followed by ground-water at 4.77 and air-water at 4.54, while air-air cannot yet be ranked on average SCOP from this corpus because the type aggregation reports no average SCOP value for that class (type_efficiency).

That first point matters because it is both simple and easy to misread. The highest average efficiency in the current model universe belongs to water-water, not to the larger mainstream categories, and the gap between first and second place is 1.38 SCOP points, from 6.15 to 4.77 (type_efficiency). Ground-water therefore sits well ahead of the broad market average SCOP of 4.55, while air-water is almost exactly on it at 4.54 (market_index_snapshot) (type_efficiency).

The air-to-air part of the seed’s claim cannot be fully verified from the supplied type-average dataset. The corpus does report 21,065 air-air models and gives their average capacity and outdoor noise, but its average SCOP field for air-air is null, so a Europe-wide 2026 class average cannot be stated responsibly here (type_efficiency). If you want the current model universe in one place, the market index snapshot and full heat-pump catalog are the right references.

How big is the air-to-air vs air-to-water gap?

The corpus supports only one side of that comparison directly: air-water averages 4.54 SCOP in the 2026 snapshot (type_efficiency). It does not provide an average SCOP for air-air, so the absolute SCOP gap and the percentage gap versus air-water cannot be calculated from the supplied data (type_efficiency).

What can be said is that the evidence base is uneven across types. The air-to-water heat-pump catalog is large enough to support a stable class average, with 30,452 models listed (type_efficiency). The air-to-air catalog slice is also large at 21,065 models, but its type-level SCOP average is absent in this extract, which blocks any clean Europe-wide comparison (type_efficiency).

That means the headline “air-to-air now beats air-to-water on efficiency” is not demonstrable from this JSON alone. The stronger, corpus-backed story is different: the 2026 hierarchy among the types with reported averages is led by water-water, then ground-water, then air-water, and the market’s biggest category is not the most efficient one (type_efficiency) (market_index_snapshot).

What the market snapshot says about model counts and type mix

The model mix is dominated by air-water and air-air, not by the efficiency leaders. Out of 60,989 total listed models, air-water accounts for 30,452, or about 49.9% of the universe; air-air accounts for 21,065, or about 34.5%; hp-water-heater for 9,228, or about 15.1%; ground-water for 213, or about 0.35%; and water-water for just 31, or about 0.05% (market_index_snapshot).

So the most efficient type on average, water-water at 6.15 SCOP, is also the smallest by model count at 31 listings (type_efficiency). Ground-water, the second-most-efficient type on average at 4.77 SCOP, has only 213 listed models (type_efficiency). By contrast, air-water is the largest category with 30,452 models despite ranking behind both water-water and ground-water on average efficiency (type_efficiency).

This is the central divergence in the 2026 snapshot: product abundance does not mirror the efficiency ranking. Buyers browsing the full catalog or the leaderboard hub are seeing a market shaped by installability and mainstream use cases as much as by headline SCOP.

What the top-scoring models reveal about the ceiling by type

Among the top 15 models by SCOP overall, 10 are air-water, 4 are water-water, and 1 is ground-water; there are zero air-air entries in the supplied top-15 overall list (top_models). That distribution is almost the inverse of the average-efficiency ranking: the rarest classes dominate the average table, but air-water dominates the very top of the all-model leaderboard because it has far more shots on goal (type_efficiency) (top_models).

The highest SCOP observed by type in the supplied leaderboards is 7.0 for air-water, held by Risch Kälte- und Klimatechnik GmbH OH I 4esr TWW W/W (top_models). Water-water peaks at 6.97, seen on Waterkotte GmbH CTC EcoTouch 525 (water/water) and Waterkotte GmbH EcoTouch DS 5034.5 T (water/water) (top_models). Ground-water peaks at 5.97 in the supplied ground-source leaderboard (top_models). For air-air, no top-SCOP models are returned in the supplied type-specific query, so the highest observed SCOP for that type cannot be stated from this corpus (top_models).

For readers comparing ceilings rather than averages, the overall top-SCOP leaderboard, air-to-water SCOP leaderboard, and ground-source SCOP leaderboard show how concentrated those high scores are.

Why efficiency alone does not settle the buyer decision

The efficiency leaders also look different on practical metrics. Water-water units average 35.65 kW and 42.0 dB outdoor noise, making them by far the largest and quietest class in this dataset (type_efficiency). Ground-water averages 18.45 kW and 58.8 dB, while air-water averages 11.83 kW and 59.8 dB (type_efficiency). Air-air, although missing a class-average SCOP here, averages just 5.41 kW and a higher 64.1 dB outdoor noise (type_efficiency).

That means “more efficient” does not necessarily mean “closer substitute” for another class. The top average-efficiency category, water-water, is not just rarer; it is also much larger on average than air-water, by 23.82 kW, and quieter by 17.8 dB (type_efficiency). Ground-water is also materially larger than air-water, by 6.62 kW, while being 1.0 dB quieter on average (type_efficiency). Air-air sits at less than half the average capacity of air-water, 5.41 kW versus 11.83 kW, and is 4.3 dB louder outdoors on average (type_efficiency).

So the buyer mistake in 2026 is to treat type labels as if they were interchangeable efficiency wrappers. They are not. Capacity, noise, and likely application envelope differ materially by class in this snapshot. For a broader sense of how these trade-offs sit in the market, see the methodology notes, the quietest-model leaderboard, and the sizing calculator.

Sources

type_efficiency — EPREL Public API · type aggregation. Snapshot: 2026-05-30.
market_index_snapshot — Househeating Pulse · Market Index v1, computed from EPREL Public API. Snapshot: 2026-05-30.
top_models — EPREL Public API via Househeating Pulse catalog. Snapshot: 2026-05-30.