What would a MetroPCS engorged with T-Mobile’s assets look like?

Last week Bloomberg reported that MetroPCS was the frontrunner to buy up any network and spectrum assets that AT&T and T-Mobile would part with in order to push their merger through. Though Bloomberg didn’t name its sources, such a deal should hardly be a surprise. From the beginning Metro has said that it was always on the lookout for cheap spectrum acquisitions and views any AT&T-T-Mobile divestiture as a prime target (CP: MetroPCS unopposed to a combined AT&T, T-Mobile).

AT&T’s proposed acquisition is looking less and less likely as different regulators weigh in. The U.S. Department of Justice has filed an antitrust suit (CP: DOJ lawsuit dashes AT&T’s hopes of an easy merger review), which several states have joined, and the signals from the FCC look increasingly grim for AT&T’s hopes (Unfiltered: FCC gets tough with AT&T on merger job claims).

Divesting markets, subscribers and spectrum may seem like a good way to ease regulators’ concerns, but the DOJ’s primary beef with the deal seems to be the elimination of T-Mobile as a low cost nationwide competitor in the market. Giving up assets won’t satisfy that problem. But let’s assume that AT&T is making progress and it sells a portion of assets to MetroPCS, what impact would that have on Metro?

AT&T and T-Mobile have a lot of one thing that MetroPCS wants: spectrum—and not just any old spectrum, but frequencies in the PCS and Advanced Wireless Services (AWS) bands over which Metro has built its CDMA and LTE networks. One of AT&T’s main justifications for the merger is consolidation in AWS, creating a huge LTE pipe for its future mobile broadband services. AT&T probably won’t be willing to sacrifice any of its own of T-Mobile’s AWS spectrum so that leaves PCS.

If MetroPCS is looking merely to expand capacity within its current footprint or build around the edges of the current markets, then the transaction is fairly straightforward. It will take T-Mobile’s AWS spectrum and be done with it. Metro has stated its future growth will be through LTE—it’s already moving SMS to the all-IP network and plans to eventually do the same with voice—so it would immediately designate any new 10 MHz or 20 MHz PCS license to expanding its now-limited LTE capacity.

But if MetroPCS wants to use the opportunity to expand its footprint—which it’s demonstrated every ambition of doing—then things become immensely more complicated. The Bloomberg story states that both subscribers and spectrum are likely part of deal being negotiated between the three parties. Subscribers would mean Metro is buying a network, and that means GSM.

Though MetroPCS is a CDMA operator, it probably wouldn’t be opposed to running a GSM network in different territories, at least temporarily. As a regional provider with a far flung footprint, it’s accustomed to thinking of its network as market pieces than a continuous whole. And any sale of GSM assets to Metro would certainly be accompanied by a reasonable GSM roaming agreement with the emerging AT&T juggernaut so Metro wouldn’t have to worry about its new subscribers being stranded or provisioning any crazy tri-mode phones.

MetroPCS, however, would only run that GSM network long enough to bring LTE online and transition its customers over to the new network. As pointed out earlier, MetroPCS is being very aggressive when it comes to Voice-over-LTE and IMS. As soon as it can get affordable LTE handsets for all of its customers it will look to shut down as many legacy networks as possible. GSM would be first on the list.

The problem is MetroPCS simply can’t go out and buy T-Mobile’s GSM network in any given market. AT&T wants to keep the high-speed packet access plus (HSPA+) portion of that network so it can eventually have that spectrum for LTE. That means AT&T will have to decouple the GSM network from the HSPA network, which is easier said than done. Those two networks share cell sites, backhaul, site power and housing and are linked into the same data and voice cores. They may be separate radio networks with separate base stations and antenna, but T-Mobile, like any good operator should, is splitting whatever other resources it can between them.

If both AT&T and MetroPCS were planning on maintaining those two networks for any length of time, then they’d face an impossible situation—one would have to replicate a good deal of the other’s infrastructure and deployment to create two stand alone and viable networks. But the beauty of the AT&T and Metro situation is neither one wants to maintain the status quo for very long. While Metro would want to shut down GSM in favor of LTE, AT&T wants to shut down HSPA+ to re-farm its AWS spectrum for LTE. Both moves necessitate brand new networks.

The only glitch is neither operator is ready to take that plunge immediately. If Metro could move entirely to LTE today, it wouldn’t bother with the GSM networks and would simply scoop up the spectrum. AT&T can’t just shut down HSPA+ without leaving millions of T-Mobile subs without mobile broadband, nor can it simply shift those subscribers over to its own HSPA+ network at PCS without inviting catastrophe. So sunsetting T-Mobile’s HSPA+ service to facilitate a network transfer to Metro isn’t an option either.

The two operators would either need to engage in some bizarre network-sharing plan or, more likely, have one lease the equipment capacity back to the other. AT&T would sell the PCS spectrum to MetroPCS outright, which from AT&T’s point of view would satisfy its divestiture requirement. Then AT&T would either lease the GSM equipment and backhaul capacity to MetroPCS or would sell the whole network to Metro for a nominal price and lease the HSPA equipment back from Metro. My bet is Metro would want the latter. Why buy a network that you’re only going to junk in a few years anyway. Plus AT&T may be able to reuse that HSPA+ gear, retuning it for PCS of cellular.