Technology adoption is one the most important elements of a firm's strategy. In this paper, we address an essential, yet largely overlooked, question: What should a firm do when faced with several alternative proprietary designs of a new technology? In our base case we assume there are two technology designs, each described by an independent stochastic process of technology evolution. We show that, in equilibrium, a buyer chooses the leading technology design as soon as the discounted payoff from doing so is positive. When the option value of waiting is very high, it is jointly optimal to delay adoption. But because sellers cannot commit not to extract all of the buyer's future rents, inefficiently early adoption takes place. Strategies that improve commitment to low future license fees, such as increasing the number of competitors or cross-licensing, may alleviate the hold up problem. Although previous research stressed the benefit of such commitments in terms of increasing the rate of technology adoption, we present a class of cases when the benefit from commitment is efficiently to delay adoption.