The choice between high voltage AC (HVAC) and HVDC export in offshore wind is often made by comparing project economics at fixed offshore substation (OSS) locations, typically close to the array centroid. We argue this is ill-posed: HVAC and HVDC have different cost-minimising OSS positions, so a fixed-location comparison can return a locally optimal but globally wrong technology answer.

We demonstrate this on a representative 2.1 GW offshore test case built from public data. Joint OSS siting and inter-array cable routing is formulated as a mixed-integer linear program that captures spatial installation constraints, electrical losses, expected energy not transmitted, and lifetime project economics. Wind turbine positions and the point of common coupling are fixed, while OSS position is the decision variable.

Four design cases combining HVAC/HVDC export and 66/132 kV inter-array systems are evaluated. For each candidate OSS location, the model optimises inter-array and export cable routing and computes the associated lifetime cost, producing a cost surface as a function of OSS position.

The four cost-minimising OSS positions do not coincide. HVAC favours 31–37 km from the point of common coupling; HVDC favours 45–48 km, deeper into the concession. The resulting HVAC and HVDC optima are 9–17 km apart, while inter-array voltage introduces smaller secondary shifts. Export technology dominates OSS position and inter-array voltage is secondary, but both shift the optimum. The asymmetry is physically interpretable: HVAC export cost grows steeply with distance due to capacitive charging current, pulling the OSS shorewards; HVDC is length-insensitive, so the OSS sits closer to the wind farm centroid to minimise inter-array length.

The position-induced cost spread is highly asymmetric: HVAC's surface varies several times more than HVDC's. The two total-cost ranges overlap, so the technology answer depends on where the OSS is sited; HVAC wins at HVAC's optimum, HVDC wins if the OSS is placed toward HVDC's. The reversal occurs within the position space evaluated for this site, not only at hypothetical marginal sites.

The contribution is methodological: joint optimisation of export technology, inter-array cable voltage and OSS position can yield different technology selections than the standard fixed-location comparison, with cost consequences of hundreds of millions of euros at GW scale. The proposed cost-surface approach is a transferable, tractable design tool for early-stage screening, generalising across array sizes and shore distances.