AT&T's HSPA+ was pretty bad, even in comparison to T-Mobile's HSPA+. Even if you were in a place that got no LTE signal, you'd get better performance on T-Mobil

AT&T's HSPA+ was pretty bad, even in comparison to T-Mobile's HSPA+. Even if you were in a place that got no LTE signal, you'd get better performance on T-Mobile than AT&T. AT&T was stuck on 21 Mbps vanilla HSPA+, while T-Mobile rolled out 42 Mbps DC-HSPA+ in 2011 (with a surprise midnight announcement that they'd already converted a huge chunk of their network and just turned it on). DC-HSPA+ had real-world performance better than or equal to AT&T and Verizon's LTE networks at the time, which were allegedly 100 Mbps but didn't perform like it, so I honestly don't blame T-Mobile for marketing it as being equal to Verizon's network. For context, this was back when AT&T was trying to buy T-Mobile, and the reason why T-Mobile was enthusiastic about it was because they were out of spectrum and didn't have any room to roll out LTE. So instead they went all-out in making their HSPA+ network as fantastic as possible until they could get bought by somebody with spectrum (i.e. AT&T). Rolling out DC-HSPA+ ahead of schedule was just the start; they had plans to follow that up with MIMO bringing speeds up to 84 Mbps. But then government interference caused the deal to fall through, which triggered a trap clause that forced AT&T to cede a decent chunk of spectrum to T-Mobile, so they scrapped their MIMO plans and rolled out LTE on their newly-acquired spectrum. And they very much had an incentive to make their LTE network fantastic, because they needed it to perform better than their DC-HSPA+ network to be worth it for people to buy new phones. Being newer than AT&T's and Verizon's LTE networks (and maybe Sprint's, but I can't remember if Sprint switched to LTE before or after this) probably helped them, as they got to start it out with all-new hardware.