Abstract This study sheds new light on the variability of wind power across the Australian NEM (National Electricity Market) and in doing so gives an insight on the potential network configuration for a high RE (Renewable Electricity) future. We present idealised cost-minimised simulations for the NEM utilising onshore wind, large-scale solar, pumped hydro and OCGT (open cycle gas turbines) technologies. A model using gridded meteorological data from the regional ACCESS-R (Australian Community Climate and Earth-System Simulator) simulates wind and solar technology output along with generation from OCGT to meet demand in the NEM for the period 2010–2011. A cost for connecting each location to the nearest major load centre is introduced and a base scenario created from an initial connection cost of $1 M/km. A sensitivity study reveals that a cost of $8 M/km results in the contraction of all renewable resources to four major wind installations. Compared to the base scenario the four major wind locations share much of the variability in renewable energy output, demonstrating that the NEM region has four distinct wind regimes. Separated by 1,400 km these four wind installations provide an optimisation-based decorrelation length for the NEM. This information is particularly useful for long-term planners of large-scale energy infrastructure.