Microsoft Research has some interesting recent work on using FPGAs for accelerating the convolution steps of DNN training at least: https://www.microsoft.com/en

Microsoft Research has some interesting recent work on using FPGAs for accelerating the convolution steps of DNN training at least: https://www.microsoft.com/en-us/research/publication/accelerating-deep-convolutional-neural-networks-using-specialized-hardware/?from=http%3A%2F%2Fresearch.microsoft.com%2Fpubs%2F240715%2Fcnn%2520whitepaper.pdf https://www.microsoft.com/en-us/research/publication/acceler... The big win for them is a 10x reduction in power usage, since in a datacenter/cloud environment this is more important. Still at the research stage though.