Here, have a proof:
Take a look at the access times. They're as low as 0.09ms and that includes the SSD's overhead. So, the true USB latency is actually way below 0.1ms. Do you really think you can hear the difference? Scientific research has shown that human brain does not perceive a delay of 10ms as a separate sound. In practice, you need something like 20ms to actually hear the difference, and when playing an instrument live, you can "feel" the difference when the latency is slightly above 10ms. But then, what are we comparing? 0.1ms to 10ms? That is 100 times difference. So USB latency is 100 times lower than you need it.
The problem is not in USB itself. The problem is in how your interface implements it, and that can be fixed by just getting a better interface. But USB is NOT the culprit here.
Also, if you take a look over here:
People are reporting as low as 30 MICROseconds latency for a round trip. If you only have an output, no input, that leaves us with 15us = 0.015ms latency.