Ship developers went at times to great lengths to increase warship speeds. A single knot more or less required a dramatic change of installed power, though: For example, the Iowa (BB61) class improved the armour protection only a bit over their predecessors the North Carolina battleships (BB55).
This and the increase of displacement were largely compensated for by the vastly improved hydrodynamic shape of the hull. The requirement that drove the installed power upwards by 75% was the increase of speed from about 28 kts to about 33 kts. The much-enlarged boiler and turbine rooms drove the displacement and costs up by a lot, of course.
As so often, the obvious specs got all the attention, whereas the hidden values of navies got little.
The following quote should make obvious why certain utterly un-sexy tasks may in wartime actually trump the expense and attention-grabbing boasting of peacetime:
A most complete towing test of hull resistance was made on the Japanese ex-destroyer Yudachi. This 234-foot vessel was docked, painted, and had the propeller removed in March, 1931. Immediately after undocking it as subjected to systematic towing tests which were repeated at intervals to show the effect of fouling. The results of the tests on the Yudachi [...] demonstrate the very great increase in resistance which developed while the ship remained at anchor. [...] In 375 days the total resistance is exactly doubled. [...] In Figure 3, the loss in speed with a towing force of 10 tons is plotted against the time at mooring. This force produced a speed of 20 knots with the freshly painted hull. After 375 days the speed had fallen to 15.4 knots [...].
Developers, bureaucrats, politicians and naval enthusiasts accepted great expenses in exchange for a single or few more knots top speed. Yet during actual task force movements the slowest ship always dictated the pace and thorough maintenance was often times much more important than spec sheet glory.
related: The price of speed at sea