DDN: Parallelizes Big Data:The revolution of deep learning is that suddenly we can make machine learning cleverer by giving it more data. This coincides with the advent of GPUs, which could parallelize this problem. And then all of a sudden, we have not only a big a potential software algorithm, a framework that we can use to bring new competitive value, make decisions faster, quicker, easier, more autonomous for these organizations. But there's a vehicle to do it in GPUs, a highly new parallelized way of doing it. And so effectively, all these organizations with any reasonable metadata data are playing A.I., whether it's that small scale or at large scale. And these customers tend to be working with video data, images and with audio. So streams of large volumes of data where they're trying to extract value of that in a huge variety of ways.