AT&T isn't in as nice a spectrum position as Verizon, but the situation isn't what you've portrayed. AT&T has AWS spectrum in most markets, but it isn't a 40MHz chunk. That's still additional capacity. AT&T is also re-farming spectrum previously used for 2G/3G services for LTE. Likewise, the 150Mbps theoretical speed won't happen in real-world conditions (just as 10x10 LTE doesn't give you 75Mbps speeds). It will offer better speeds, but quoting theoretical speeds in wireless is somewhat useless.

Verizon is in a wonderful spectrum position. The 40MHz chunks that Verizon has in many markets is great. AT&T is going to have to do carrier aggregation and we'll see how well carrier aggregation works vs. larger contiguous channels. But overall Verizon doesn't have too much more spectrum than AT&T. If AT&T can pull off carrier aggregation well, there should be decent competition. Even before AT&T started diving into small cell sites, the Cingular/ATTWS merger had left the company with more cell sites than Verizon.

Verizon was really smart when it came to LTE. They knew that they could grab customers if they bought up the spectrum needed for LTE. Spectrum is a limited good and if they had the most/best, it would give them a durable advantage that's hard to compete against. They bought the open-access 700MHz spectrum whose open-access restrictions seem to have been watered down and while AT&T was fighting to buy T-Mobile (which wouldn't be approved because it would concentrate too many customers together and limit competition), Verizon bought the AWS spectrum from the cable companies which gave them a very powerful position to deploy LTE.

Verizon did very well with their LTE rollout, but AT&T covers the vast majority of their network. The area that's really different are the areas that AT&T never existed in prior to Verizon buying Alltel.

Verizon got a lot right, but definitely not 100% of the time. The iPhone hit AT&T hard. EDGE (2G GSM data) used a lot of spectrum for the original iPhone and put AT&T in the position of investing a lot in a dead technology. If you were Verizon with the knowledge we have today and had the opportunity to grab exclusivity on a 2G iPhone that would have you investing in 2G capacity (and using up your spectrum for several years), would you do it? I think they'd probably say yes. Ultimately, I think AT&T wishes the iPhone had come out a year later with 3G. The 2G iPhone put AT&T in the crappy position of using up a lot of spectrum that could have gone to 3G and requiring a lot of investment in 2G that could have gone to 3G. But it also really helped gain customers.

Sprint was a sad situation. They wanted to outsource their 4G network to Clearwire. You can't outsource your core competency. Sprint starved Clearwire of funds which meant that an aggressive network buildout just wasn't possible. Clearwire's high-frequency spectrum meant that a very large number of cell sites would be needed to cover an area. By contrast, AT&T and Verizon's 700MHz spectrum means that they can roll out fast and then add cell sites for capacity as needed. Sprint/Clearwire didn't have that option. I think Sprint was genuinely worried that if Clearwire succeeded, Sprint would become nothing more than a Clearwire MVNO.

Sprint lost a lot of time. It didn't want to fund Clearwire. Then it thought it was going to buy LightSquared's spectrum and dump Clearwire. Then Clearwire was going to build TD-LTE for Sprint. Then Sprint was going to build 5x5 LTE on their PCS spectrum. Then Sprint was getting bought and buying Clearwire. Then Dish started competing for Clearwire. This is just years of wasted time on direction.

I think Spark is moving slowly because it's hard to roll out such a high-frequency network in such a low-density country. On top of that, I think the permitting/approval process and the ability to get fiber to so many points was underestimated. SoftBank (majority owner of Sprint) has said that they deploy a ton of cell sites in Japan on the same high frequencies and get fiber to them all. But Sprint has been complaining of getting fiber and a lot of locations make it hard to build lots of cell sites.

But Sprint is doing a lot better recently. They're deploying 5x5 LTE on their 800 and 1900MHz spectrum. It doesn't have the capacity that the other three have, but it's a much better situation than a year ago. And Sprint is gaining customers (about 1M last quarter, though post-pay subscribers were flat). Sprint seems to be in the kind of position that T-Mobile was a year and a half ago and I think that bodes well for competition in the future. Sprint's 800MHz spectrum should offer good coverage and reliability for their LTE network. Their 2.5GHz spectrum could help with capacity and speed in cities. Their new CEO seems to price wisely and aggressively. I think Sprint is in a stable point today. Their network is a bit slower, but the difference between 8Mbps and 12Mbps or 12Mbps and 18Mbps isn't as relevant as the gap from 2Mbps to 10Mbps that was often the case.

I think 2015 will see a more competitive wireless landscape than we've seen in a long time. T-Mobile is draining customers from competitors and has announced bold intentions for coverage. Sprint is pricing aggressively and has a network that is reasonable and will hopefully get a lot better. The auction for 600MHz spectrum might be happening this year and might have rules favoring companies not named AT&T and Verizon. Two years ago, Verizon and AT&T had LTE while Sprint and T-Mobile languished. Today, the vast majority of Americans have four carriers that would service them well. Some might be better than others in different locations or for different purposes, but it's a huge difference from the era when we were wondering if T-Mobile would even launch LTE and when Sprint completely lacked direction.

Maybe you guys can answer this. With mainland Europe using GSM across the board, do they have these frequency allocation problems across providers? It seems to me like it would be in everyone's interest to have a large shared pool of frequency allocations and share the burden of maintaining a network using them. Then you significantly more bandwidth capability per tower.

It's worth noting that Sprint is also Virgin Mobile US and Boost Mobile ... which has a fairly large subscriber base and is very oversold. T-Mobile also has a few MVNOs that use their network (I use Simple Mobile)... It seems that AT&T has cut support for most MVNOs using their network, and Verizon afaik hasn't ever had any.

It's a bit of a race to the bottom in one way, but I never saw the point in paying more for AT&T or Verizon. At home I get very good coverage... the only time I'd rather have Verizon is on my annual road trip, and that's only a week or two. If you're in the northern us (Wyoming, Montana, Dakotas etc), Verizon is pretty much your only option.

I haven't enjoyed reading a comment on HN this much in a very long time. Thank you for this very informative post!

(I realize this comment doesn't add value, but what's the point of a commenting system that doesn't let one echoes gratitude with more than just an upvote?)