Regression optimization question

So I have multiple multivariate models that I am using, and my techniques, which could technically be defined as curve fitting however I am running about 1 million regressions every 1-3 seconds. My avg MSE right now on the tests I have done are between 1-7. If I run so many models, so that I am essentially continuously optimized for the current market conditions, will this be sufficient?

Sufficient for what? You haven't specified any goal so there is no way to know if what you are doing is or isn't sufficient. Other questions include, why run the regressions every 1-3 seconds? What significance does a mean squared error (MSE) of 1-7 have for this application (1-7 seems like odd residual error from a regression)? What is the application!? Will minimizing MSE actually accomplish your goal?

Offhand it sounds like your chances of overfitting are very high due to the frequency of reoptimization. But without knowing more about the types of models, how much data you are using to optimize, the frequency of the data, the number of variables in your models it isn't possible to tell.