The large turbines typically operate up to 25m/s and then shut down for safety. Really large offshore turbines operate ideally at an average of 11m/s [1]. You can also see some real data here [2].
However, what investors (or farm operators) really care about it LCOE. Levelized cost of Energy over the lifetime of the machine. It's that $50/MWh and considers the revenue (roughly proportional to v^3) and all the costs. After that it's an optimization game:
- Too far = costly maintenance, costly installation, longer cables, rougher seas - Too close = not enough wind (but "cheap" installation)
(Plus many other important factors: risk, incentives, NIMBY, access to capital, wind turbulence, etc, etc)
[0] https://goo.gl/images/NFny2a [1] http://www.homepages.ucl.ac.uk/~uceseug/Fluids2/Wind_Turbine... [2] http://www.wind-power-program.com/large_turbines.htm
1. Wind Power ~ v^3 (the power produced is proportional to the wind speed cubed. So a site with 8m/s average speed makes twice as much money as one with 6m/s!
2. Often, that sweet offshore-spot is too deep (in that the cost of monopiles [1] is prohibitive).
So the key decision is: If I can make it go just a little deeper, I can make 2x or 3x as much money. Thus floating turbines can unlock more revenue (but the cost variation is not so important).
[1] http://www.4coffshore.com/windfarms/monopiles-support-struct...