New Findings: What is the central question of this study? The aim was to determine, using the technique of agitated saline contrast echocardiography, whether exercise after 4–7 days at 5050 m would affect blood flow through intrapulmonary arteriovenous anastomoses (QIPAVA) compared with exercise at sea level. What is the main finding and its importance? Despite a significant increase in both cardiac output and pulmonary pressure during exercise at high altitude, there is very little (QIPAVA) at rest or during exercise after 4–7 days of acclimatization. Mathematical modelling suggests that bubble instability at high altitude is an unlikely explanation for the reduced (QIPAVA). Blood flow through intrapulmonary arteriovenous anastomoses (QIPAVA) is elevated during exercise at sea level (SL) and at rest in acute normobaric hypoxia. After high altitude (HA) acclimatization, resting (QIPAVA) is similar to that at SL, but it is unknown whether this is true during exercise at HA. We reasoned that exercise at HA (5050 m) would exacerbate (QIPAVA) as a result of heightened pulmonary arterial pressure. Using a supine cycle ergometer, seven healthy adults free from intracardiac shunts underwent an incremental exercise test at SL [25, 50 and 75% of SL peak oxygen consumption (VO2peak)] and at HA (25 and 50% of SL (VO2peak)). Echocardiography was used to determine cardiac output (Q) and pulmonary artery systolic pressure (PASP), and agitated saline contrast was used to determine (QIPAVA) (bubble score; 0–5). The principal findings were as follows: (i) (Q) was similar at SL rest (3.9 ± 0.47 l min−1) compared with HA rest (4.5 ± 0.49 l min−1; P = 0.382), but increased from rest during both SL and HA exercise (P < 0.001); (ii) PASP increased from SL rest (19.2 ± 0.7 mmHg) to HA rest (33.7 ± 2.8 mmHg; P = 0.001) and, compared with SL, PASP was further elevated during HA exercise (P = 0.003); (iii) (QIPAVA) was increased from SL rest (0) to HA rest (median = 1; P = 0.04) and increased from resting values during SL exercise (P < 0.05), but was unchanged during HA exercise (P = 0.91), despite significant increases in (Q) and PASP. Theoretical modelling of microbubble dissolution suggests that the lack of (QIPAVA) in response to exercise at HA is unlikely to be caused by saline contrast instability.