Hypothetical power dissipation by a tidal in‐stream energy conversion device was calculated for Admiralty Inlet, Washington, a highly energetic entrance channel to Puget Sound and currently a candidate for tidal energy development. Power dissipation was calculated for a device of a given capacity as a function of hub height above sea bottom (z), using acoustic Doppler current profiler data taken in 2007 and 2009 at seven locations throughout the Inlet. At five of the seven locations, where the tidal currents were predominantly bi‐directional, power dissipation density increased roughly as ~zγ , with the exponent γ narrowly ranging from 0.62 to 0.66, up to at least z = 18m. At two of the five sites where the water depth exceeded this substantially, power was found to increase further with z but at slower rates with γ = 0.19 and 0.42. The remaining two sites out of the seven were both close to shoreline features, and the currents deviated significantly from bi‐directionality. At one of the two the power increased at a slower rate of γ = 0.35 – 0.5, while at the other the increase was faster with γ = 0.91 – 0.96. The increase in power with height at the bi‐directional sites is faster than would be expected from the one‐seventh power law applicable to turbulent channels. However, it is still slower than the likely increase in the cost of device foundation with height, which would at a minimum scale as ~z due to the cost of materials alone, and likely scale faster in order for the installation to withstand the overturning moment exerted by the tidal current. Thus, placing a tidal device high in the water column to exploit stronger currents may not be economically attractive, given that the device operator needs to recoup the higher cost of device foundation required.