A system and method of optimizing cryptographic mining yields includes analyzing, by a cryptocurrency mining selection system, data associated with factors of interest for one or more cryptocurrencies using machine learning algorithms. Data that is determined to be predictive of the future value of newly mined tokens is used to determine which tokens will have the highest and lowest future values. Based on the predicted value of tokens in the future and the current value of those tokens for each cryptocurrency, the system outputs one or more instructions to buy tokens in cryptocurrencies predicted to increase in value, to sell tokens in cryptocurrencies predicted to decrease in value, and to instruct associated cryptocurrency mining hardware to switch to generating new tokens in one or more selected cryptocurrencies to maximize yields.