The performance limits of electrical fiber brushes are assessed by means of an expanded, refined theory. Measurements back to 1979 are in accord with the theory, including 1993 measurements on a set of 64 fiber brushes in a homopolar motor at the U.S. Navy Annapolis Laboratory, At /spl nu/=7.6 m/s velocity and j=1 MA/m/sup 2/ the electric loss was 0.026 W/A, and after a 20 h test the brushes showed no detectable wear. On the other end of the scale, electrical brushes are being developed for a maglev train system (American Maglev Tech of Florida, Inc.) to operate up to 150 m/s. According to the theory, this will be possible in the open atmosphere with dimensionless wear rates smaller than 10/sup -8/ at current densities up to a few MA/m/sup 2/. At speeds to 300 m/s, projected losses may range up to 4.3 W/A at flash temperatures at and below 100/spl deg/C. The newest fiber brushes preserve the high contact spot density at the ends of individually flexible fibers, independent of amount of wear. Correspondingly, with 10 cm long brushes, for example, the projected maglev brushes should last at least for a New York-Los Angeles return trip. Also, on account of the changed method of construction, manufacturing costs have decreased by more than a factor of ten in the past two years. Prospects are that with industrial manufacturing techniques the cost of fiber brushes, based on ampere hours conducted, will become easily competitive with monolithic graphite brushes. Pending further practical experience, it is therefore expected that: (1) metal fiber brushes will expand the range of sliding brush applicability and (2) ultimately displace graphite brushes for most of their present applications.