And it goes even further back than that: when Acorn got first silicon back for the ARM1, they were hoping for anything less than 1W so that affordable plastic packaging would be sufficient. When they booted up that processor, their multimeter on the power supply line didn't register anything. Further investigation showed that the supply line wasn't actually connected, and the processor had booted using a total of about a tenth of a watt of leakage from other components on the board.