I think it depends on what your core focus is as a startup and whether or not the partnership is helping you achieve the milestones that are paramount to your success.
I always like to begin with the goal (for both parties) of the partnership in mind. If that goal is core to the company milestones (as it should be) then it's worth considering a diversion in the roadmap to achieve. Conversely, if it's something that isn't going to move the needle toward a core milestone, then it's probably not worth adjusting the roadmap.
I've found that framing things this way really helps in discussions with the CEO and engineering and ensures that the roadmap only changes when it really needs to.