How can more computing power improve strategy performance?

I wonder if more computing power could - to some extend - have a positive influence on my strategies' results. So far, all strategies I saw do not actually require a lot of cpu. What are your thoughts about that?

Really depends on how much number crunching you are doing and how quickly you need that crunching done. The longer your number crunching takes, the more time will elapse between the appearance of a trading signal and your entry into the position. For large time frames, this additional time may not make much of a difference as it would be only a fraction of the move that is typically targeted. On a short time frame, any additional latency could be more significant. Indicator-based strategies will generally run lightning fast, but if you are training machine learning models by calling R, you can get noticeable delays in execution, particularly in a portfolio situation.