What a watch means when it says VO2 max

What VO2 max actually is

VO2 max is the maximum rate at which your body can take in and use oxygen during exercise, in milliliters per kilogram of body weight per minute. Measuring it properly means running or cycling to exhaustion while breathing into a mask that analyzes expired gas — cardiopulmonary exercise testing, done in a laboratory.

Your watch has no mask. What it has is your pace, your heart rate and a model of how those two relate in trained humans. The estimate is the output of that model.

That is not a criticism — it is a genuinely clever inference and it is free. It is just a completely different thing from a measurement, and the difference shows up in the numbers.

What the two studies found

Lambe et al. (2025) tested 28 participants across a range of Apple Watch models against indirect calorimetry on a treadmill. The watches underestimated VO2 max by 6.07 mL/kg/minon average, with a mean absolute percentage error of 13.31% and limits of agreement running from −6.11 to +18.26. The authors concluded the estimates "are not sufficiently accurate to inform clinical decision-making", while noting they may still be useful for general fitness monitoring.

Caserman et al. (2024) tested an Apple Watch Series 7 against a portable gas analyser on 19 participants. Lab measurement averaged 45.88 mL/kg/min against the watch's 41.37 — a bias of −4.51 — with a MAPE of 15.79% and an intraclass correlation of 0.47, which is poor reliability.

Two studies, two research groups, different protocols, same direction. That consistency is what makes this worth acting on.

The finding that matters most to a trained runner

The 2024 study stratified its results by fitness level, and the pattern is the useful part:

The estimate overshot in the least fit group and undershot badly in the fittest one — a 12-point gap in the excellent group, which is roughly the difference between a competent club runner and an elite one.

If you are reading this because you train seriously, that is the row that applies to you, and it says the number on your wrist is likely to be substantially low. Both studies used small samples, and both authors flag it.

Why the estimate behaves this way

The model infers aerobic capacity from how fast you can go at a given heart rate. That relationship is real, but it is affected by running economy, heat, hills, hydration, altitude and the specific sessions the watch happened to see.

At the top end, the model runs out of room: highly trained runners differ from each other in ways — economy, lactate handling, efficiency — that a pace-and-heart-rate relationship cannot resolve. That is a plausible explanation for the fitness-stratified result, and it is our reading rather than a finding either paper states.

The 2024 authors add a fairness point worth repeating: their watch figure came from a single outdoor session, where Apple's own internal validation used an average of 441 days of data per user. A watch given more data may do better than these studies suggest.

What the number is actually good for

Trends, on the same device, over months. A systematic bias that stays roughly constant still lets you see a real improvement, because you are comparing the estimate against itself rather than against a laboratory.

Not for comparing across brands. Different manufacturers use different models. Two watches on the same wrist will disagree, and neither disagreement is evidence about you.

Not as a number to quote. If it matters to you what your VO2 max actually is — for a training prescription, or a clinical reason — the published evidence says get it measured. Both studies are explicit that the watch value is not adequate for decisions of that kind.

What about Garmin, COROS and Polar?

We could not find equivalent peer-reviewed validations for their current watches. Both studies above tested Apple, which is the recurring pattern on this site: Apple gets measured, everyone else gets marketed.

It would be wrong to conclude that other brands are better or worse. It is unknown, and under our evidence hierarchy unknown is its own category rather than a quiet pass — which is why the Apple Watch Ultra 3 still scores above the brands with no data at all, even though the data it has is unflattering.