https://www.microsoft.com/en-us/research/publication/crypton... (2016)

> We demonstrate CryptoNets on the MNIST optical character recognition tasks. CryptoNets achieve 99% accuracy and can make more than 51000 predictions per hour on a single PC. Therefore, they allow high throughput, accurate, and private predictions.

It’s starting to, yes, in particular for machine learning. There is a yearly competition called iDash where people show the performances of their homomorphic schemes. This year should be very interesting