What is the optimal timing of inventory investment for a firm when its forecasts of demand and price improve with time but are correlated with the prices of traded assets in the financial markets? We consider this problem using a single period inventory model where demand is realized at time T and the stocking decision may be made at any time in the interval [0, T ]. The demand forecast evolves as a geometric Brownian motion, while the price forecase exhibits mean-reversion. The processes for the firm's value as well as that of the market evolve as geometric Brownian motions. We show that the right to make the optimal inventory decision is a modified American-style option. However, since the stochastic variables defining the forecasts of demand and price are not tradeable, we use a risk-adjusted valuation approach for incomplete markets to determine the optimal timing strategy.We derive the value of the inventory postponement option and the optimal exercise decision, that is, we present and analyze conditions under which early exercise of the inventory option takes place, and those under which postponement occurs. We also derive the comparative statics of optimal timing and stocking decisions relative to the key parameters: the market price of risk (and the cost of capital), the volatilities of price and demand, the correlation between these two variables and the return on the market portfolio, and the procurement cost.We illustrate the empirical validity of our model by testing it on firms in the gold mining industry. We show that our model performs well in explaining the variation in firm inventory as well as gross profit, and provides new evidence on the impact of price uncertainty on these variables.