If I understand correctly, the float data type (which defaults to float (53)) can only store 17 digits.

Whereas '9 to the power 28' requires 27 digits.

So must I assume that the trailing 10 digits are just simply truncated off, and this is what leads to the inaccuracy?

So instead of working with

523347633027360537213511521

mssql is working with

523347633027360530000000000 ?