Aim It is an ongoing discussion the extent to which oxygen delivery and oxygen extraction contribute to an increased muscle oxygen uptake during dynamic exercise. It has been proposed that local muscle factors including the capillary bed and mitochondrial oxidative capacity play a large role in prolonged low‐intensity training of a small muscle group when the cardiac output capacity is not directly limiting. The purpose of this study was to investigate the relative roles of circulatory and muscle metabolic mechanisms by which prolonged low‐intensity exercise training alters regional muscle VO2. Methods In nine healthy volunteers (seven males, two females), haemodynamic and metabolic responses to incremental arm cycling were measured by the Fick method and biopsy of the deltoid and triceps muscles before and after 42 days of skiing for 6 h day−1 at 60% max heart rate. Results Peak pulmonary VO2 during arm crank was unchanged after training (2.38 ± 0.19 vs. 2.18 ± 0.2 L min−1 pre‐training) yet arm VO2 (1.04 ± 0.08 vs. 0.83 ± 0.1 L min1, P < 0.05) and power output (137 ± 9 vs. 114 ± 10 Watts) were increased along with a higher arm blood flow (7.9 ± 0.5 vs. 6.8 ± 0.6 L min−1, P < 0.05) and expanded muscle capillary volume (76 ± 7 vs. 62 ± 4 mL, P < 0.05). Muscle O2 diffusion capacity (16.2 ± 1 vs. 12.5 ± 0.9 mL min−1 mHg−1, P < 0.05) and O2 extraction (68 ± 1 vs. 62 ± 1%, P < 0.05) were enhanced at a similar mean capillary transit time (569 ± 43 vs. 564 ± 31 ms) and P50 (35.8 ± 0.7 vs. 35 ± 0.8), whereas mitochondrial O2 flux capacity was unchanged (147 ± 6 mL kg min−1 vs. 146 ± 8 mL kg min−1). Conclusion The mechanisms underlying the increase in peak arm VO2 with prolonged low‐intensity training in previously untrained subjects are an increased convective O2 delivery specifically to the muscles of the arm combined with a larger capillary–muscle surface area that enhance diffusional O2 conductance, with no apparent role of mitochondrial respiratory capacity.