I review the main theoretical conclusions from high‐energy spectral observations of Galactic Black Hole Candidate systems. With an eye towards future missions, I show how the ability to reach such conclusions depends critically on the type of spectral coverage one has available. Perhaps the most important factor in determining how well a set of observations constrains theoretical models is the range of energies that can be observed in a simultaneous manner, the wider the better. High spectral resolution observations over a narrow energy band, e.g., as provided by Chandra and eventually ASTRO E‐2, are potentially extremely interesting but not as useful as one might first think without an understanding of the underlying broad‐band continuum. A snapshot, time‐integrated spectrum that has very wide energy coverage, however, is still not sufficient to break all the current model degeneracies. The tightest model constraints instead come from combining spectral with timing information. The RXTE satellite excelled in this regard, enabling unprecedented spectral variability studies on both short and long timescales by virtue of its large collection area, broad energy response, and flexible scheduling. Future instruments, or combinations of instruments, must strive for a similar balance of timing and broad‐band spectral capabilities if they are to prove as effective as RXTE. © 2004 American Institute of Physics