There would definitely be a first mover advantage. On the other hand, for most other VLSI applications, improved energy usage is a PERMANENT advantage: If e.g. Apple improves energy efficiency of Annn chips, it enables them to build better devices, even if the rest of the industry catches up.

For Bitcoin, my understanding is that this is not the case, because any cost improvement in the industry will attract more miners until difficulty adjusts to the previous equilibrium where (cost of mining a bitcoin) = (value of a bitcoin)-(some risk premium)