At some point the hyperscalers end up decommissioning their servers because of perf/watt concerns. I don't think they bother with single failures for the reason

At some point the hyperscalers end up decommissioning their servers because of perf/watt concerns. I don't think they bother with single failures for the reasons you've stated, but they do decommission full sections of DCs once the wattage calculus no longer makes sense.