Chips will get better, but we're approaching the limit of Moore's Law in terms of shrinking them to get the exponential gains we've seen over the past half cent

Chips will get better, but we're approaching the limit of Moore's Law in terms of shrinking them to get the exponential gains we've seen over the past half century. The improvements in Hash/W are going to continue to slow until they're not worth the additional cost to fabricate. (Unless some radically new transistor technologies arise) My assumptions about the cost of mining were assuming that one can count the cost of the mining chip as negligible, which is certainly not the case now, but I believe will be the case in future as the chips get much cheaper, much more abundant and the period between iterations of the chips increases. Industrial miners will use the most efficient chips on the market, but their primary cost will be the electricity.