This paper investigates the effects of having a separate innovation unit on exploration, exploitation, and ambidexterity in manufacturing and service firms according to traditional paradigms of the innovation management discipline that innovation units should be organized in a separate department. Many manufacturing firms have such a unit while few service firms do. This paper sets out to investigate the advantages of having such a unit for exploration, exploitation, and ambidexterity, and whether there are differences between manufacturing and service firms that could help explain why such units are present or absent. The literature suggests that a separate innovation unit has a positive effect on exploration and ambidexterity in manufacturing firms. However, the effect on improving operational activities, that is, exploitation, is unclear. If exploration and exploitation are two ends of a continuum, as the literature suggests, more exploration comes at the cost of exploitation. On the other hand, others have suggested the possibility of an orthogonal relationship, where a separate unit can simultaneously enhance exploration and exploitation. In this paper, the Dutch Community Innovation Survey (CIS) is used to investigate these relationships for manufacturing and service firms, with a question added to the survey regarding the locus of innovation within each firm, that is, mostly within a dedicated innovation unit or dispersed throughout the firm. Our findings show that a separate innovation unit increases exploration, exploitation, and ambidexterity in both manufacturing and service firms. It thereby provides support for the orthogonal view of ambidexterity. A separate unit enhances the ability to exploit and be ambidextrous equally in service and manufacturing firms, but has a weaker positive effect on exploration, and exploratory and ambidextrous performance in service firms. This finding implies that both manufacturing and service firms benefit from having a separate innovation unit, with the advantages being greatest for manufacturing firms. In service firms, such an innovation unit alone may not be sufficient, as such units are expensive to maintain, while they contribute less to ambidextrous performance than in manufacturing firms. Based on the latter finding, future studies should make a distinction between the ability to be ambidextrous in creating exploratory and exploitative innovations, and ambidextrous performance, the ability to gain financially from engaging in both types of activities simultaneously.