At ethOS, we try to keep up with market conditions that might affect our core competencies and value propositions. So, we decided to run some analysis to see if ASICs represented a meaningful percentage of the Ethereum network hashrate.Our analysis concluded that they do not.
Our front page shows a number that represents the total number of rigs running on ethOS. At this time, it is slightly over 50,000. During the peak, it was 120,000. Our statistics page shows that over 90% of our customers mine Ethereum (it has never fallen under 90%). We have data going back two years for these numbers. We looked at the network hashrate and were able to determine that ethOS rigs consistently represent a 7% contribution to the network hashrate.
Therefore, we can reasonably say that ethOS is the best publically available sample of GPU rigs, compared to any other possible polls, Twitter feedback, social media analysis, etc. So, we took our data of two years, isolated to only rigs mining Ethereum, and ran a correlation of this data to the network hashrate.
We found a correlation of 0.99.
This means that as our customers turned off GPU rigs, the network hashrate fell proportionately. Last month, we briefly saw an increase to 60,000 ethOS rigs and saw the same increase in network hashrate.
Since it is not possible to run our software on ASICs, we can reasonably determine that ASICs do not contribute a meaningful amount to the network hashrate. If something other than GPUs were mining Ethereum, we would not see such a strong correlation. We've seen statements like "ASICs on the Ethereum network are 70%, 50%, 30%, 25%, (et cetera) of the hashrate", and our analysis cannot determine the prevalence of ASICs to be anywhere near those percentages.