Purpose <p>Assessing the validity of maximum oxygen uptake (V̇O₂_max) estimates provided by a commercially available smartwatch (Garmin Forerunner 245, Garmin Ltd., Olathe, USA) compared to laboratory-based respiratory gas analysis in moderately-to-highly trained athletes.</p> Methods <p>Thirty-five endurance athletes (Tier 2–3 athletes, 24 males, 11 females; age: 25.1 ± 3.5 years; V̇O₂_max: 60.1 ± 8.2 ml·min⁻¹·kg⁻¹) completed a treadmill ramp test with respiratory gas analysis to determine criterion V̇O₂_max. Additionally, each athlete performed two submaximal 15-min outdoor runs at &gt; 70% of maximum heart rate, during which the smartwatch estimated V̇O₂_max. Athletes were stratified into moderately trained (V̇O₂_max ≤ 59.8 ml·min⁻¹·kg⁻¹) and highly trained (V̇O₂_max &gt; 59.8 ml·min⁻¹·kg⁻¹) subgroups.</p> Results <p>Across all athletes, the smartwatch underestimated V̇O₂_max [mean differences: − 4.73 ml·min⁻¹·kg⁻¹ (run 1), -4.05 ml·min⁻¹·kg⁻¹ (run 2)]. Intraclass correlation coefficients (ICC) indicated moderate agreement between smartwatch and laboratory values (run 1: ICC = 0.71 [95% CI: 0.03–0.90]; run 2: ICC = 0.75 [95% CI: 0.17–0.91]), with mean absolute percentage errors (MAPE) of 7.9% and 7.2%. Subgroup analyses revealed better accuracy of smartwatch estimated V̇O₂_max in moderately trained group (MAPE: 4.1–2.8%; ICC: 0.63–0.66 [95% CI: 0.09–0.87]), whereas in highly trained athletes, the smartwatch underestimated V̇O₂_max by 6.3 ml·min⁻¹·kg⁻¹ (MAPE: 10.4–9.4%; ICC: 0.34–0.41 [95% CI: − 0.11–0.75]).</p> Conclusion <p>Smartwatch-derived V̇O₂_max estimates are valid in moderately trained athletes but less valid in highly trained individuals. While smartwatches are useful for general monitoring, caution is warranted in their interpretation, particularly in highly trained individuals. Laboratory-based gas analysis remains the preferred method when precision is required.</p>