We analyse the effect of commodity price cycles on firm investment decisions at the project level, by considering the decision to delay, cancel or complete a project as initially announced. In particular, we use logit and duration models of competing risks on a novel dataset of announced investment projects in Peru from different economic sectors. The empirical framework for the timing of investment is motivated by real option models for projects that take time to build, with commodity prices used as a proxy of expected future income and their volatility as a proxy for uncertainty.Our results suggest that both a reduction in commodity prices and an increase in volatility increase the probability to delay investment in the mining sector, with an amplification effect when both simultaneously occur. In other sectors, delays in implementation occur more often in periods of high volatility. Probability regressions under a competing risk framework suggest that higher commodity prices lead to a higher probability of completion in all sectors of the economy.