> Saw some good news though - last year internet speeds increased nearly 40% in the US.

This is within the "Moore's law applied to network hardware" range and is not unusual.

It's also somewhat misleading because of how they're measuring (mean download speed), because customers have recently been abandoning DSL for (faster) cable. But if a 10Mbps DSL customer switches to 20Mbps cable, it shows up in the numbers as a 100% speed increase even if the cable network is no faster than it was before, or is faster than before but by significantly less than the 100% they're getting counted as.

In other words, it's claiming the destruction of competition as a consumer benefit because the market share loss of the dying competitor brings up the average.