We show that the incumbent's investment in R&D knowledge is strictly decreasing for levels of complementary assets availability at which a start-up does not form and it remains constant, although inefficiently provided, for higher degrees of complementary assets availability. One may be led to think that the incumbent's incentives to invest in R&D when the availability of complementary assets is such that a start-up form are always lower than those when the employees remains with the firm. Instead, our main result shows that moving from a scenario without start-up formation to one with start-up formation may as well induce an increase in the optimal incumbent's investment in R&D(consistently with the empirical evidence reported above for the computer game industry). This occurs if the costs of start-up formation grow slower than market profits as the incumbent's R&D investment increases. To gain an economic intuition, note that the joint surplus in the bargaining between the incumbent and the employee is positively related to the costs of start-up formation. Hence, if the marginal effect of R&D investment on the costs of start-up formation is smaller than that on market profits, the incumbent's marginal returns from investment are lower if its payoff depends on the joint bargaining surplus(i.e. if the employee stays in the firm) rather than on market profits(i.e. if a start-up forms). Importantly, our main result still holds when allowing for the possibility of setting ex ante wages, and for that of writing ex ante contingent contracts