It's an analog system. Which means that accuracy is naturally limited. However a single analog math operation requires the same energy as a single bit flip in

It's an analog system. Which means that accuracy is naturally limited. However a single analog math operation requires the same energy as a single bit flip in a digital computer. And it takes a lot of bit flips to do a single floating point operation. So a digital calculation can be approximated with far less energy and hardware. And neural nets don't need digital precision to produce useful results.