[johnnygood]

That's quite an accusation. It's incredibly unlikely that PC Magazine is risking the FTC going after it by publishing something paid for by AT&T without disclosing it. Frankly, that was an unjustified attack to discredit without providing a compelling counter-argument or other evidence.

Moreover, 40Mbps isn't unrealistic on an empty network for an absolute maximum. Average speeds of 22Mbps seem realistic given Verizon's 16Mbps performance and the fact that Verizon probably has a couple million customers on its LTE network. Likewise, Verizon's absolute maximum is 50% more than their average speed. On an empty network, I'd expect more divergence in speed from average to maximum making AT&T's ~85% difference believable. Likewise, if you visit wireless forums, you'll find plenty of Verizon users posting their maximum rates over LTE and seeing similar speeds.

And these tests were being done using Speedtest.net run by Ookla and endorsed by the FCC. While I don't know the specifics of how it works, it is the standard in testing broadband speed. Without evidence that it can be tampered with, I'd listen to the FCC.

The real question is how AT&T and Verizon's networks will hold up as more customers come on board. Last quarter, Verizon sold only 1.2M LTE devices. AT&T has just started selling LTE devices. We might see the speeds of both networks lowered as the number of customers on the networks increases.

Many people dislike AT&T. PC Magazine is testing an unloaded AT&T LTE network and a mostly unloaded Verizon LTE network. These numbers seem very reasonable given those conditions. It doesn't mean that AT&T will remain at these speeds or remain "better" than Verizon (if your purpose is to defend a cherished carrier). It's simply an initial test of an unloaded network that shows promise for the network. Whether that promise will be realized or not we'll have to wait and see.