What is the central question of this study? The pathophysiology of acute mountain sickness (AMS), involving the respiratory, renal and cerebrovascular systems, remains poorly understood. How do the early adaptations in these systems during a simulated altitude of 5000 m relate to AMS risk? What is the main finding and its importance? The rate of blood alkalosis and cerebral artery dilatation predict AMS severity during the first 10 h of exposure to a simulated altitude of 5000 m. Slow metabolic compensation by the kidneys of respiratory alkalosis attributable to a brisk breathing response together with excessive brain blood vessel dilatation might be involved in early development of AMS. The complex pathophysiology of acute mountain sickness (AMS) remains poorly understood and is likely to involve maladaptive responses of the respiratory, renal and cerebrovascular systems to hypoxia. Using stepwise linear regression, we tested the hypothesis that exacerbated respiratory alkalosis, as a result of a brisk ventilatory response, sluggish renal compensation in acute hypoxia and dysregulation of cerebral perfusion predict AMS severity. We assessed the Lake Louise score (LLS, an index of AMS severity), fluid balance, ventilation, venous pH, bicarbonate, sodium and creatinine concentrations, body weight, urinary pH and cerebral blood flow [internal carotid artery (ICA) and vertebral artery (VA) blood flow and diameter], in 27 healthy individuals (13 women) throughout 10 h exposures to normobaric normoxia (fraction of inspired O2 = 0.21) and normobaric hypoxia (fraction of inspired O2 = 0.117, simulated 5000 m) in a randomized, single-blinded manner. In comparison to normoxia, hypoxia increased the LLS, ventilation, venous and urinary pH, and blood flow and diameter in the ICA and VA, while venous concentrations of both bicarbonate and creatinine were decreased (P < 0.001 for all). There were significant correlations between AMS severity and the rates of change in blood pH, sodium concentration and VA diameter and more positive fluid balance (P < 0.05). Stepwise regression found increased blood pH [beta coefficient (β) = 0.589, P < 0.001] and VA diameter (β = 0.418, P = 0.008) to be significant predictors of AMS severity in our cohort [F(2, 20) = 16.1, R2  = 0.617, P < 0.001, n = 24], accounting for 62% of the variance in peak LLS. Using classic regression variable selection, our data implicate the degree of respiratory alkalosis and cerebrovascular dilatation in the early stages of AMS development.