— This is the second part of the article. You can catch up on the first part here.
T-Mobile’s campaign promise
While Verizon has to draw enough customers for each small cell site to recoup its costs, T-Mobile may be setting itself up for the opposite problem—too many subscribers downloading too much data on its promised home-5G service.
The self-styled “uncarrier”—having called millimeter-wave 5G unworkable outside of dense urban pockets—has folded a different approach to 5G home broadband into its campaign to merge with Sprint.
Let us consummate this deal, T-Mobile told regulators, and we’ll bring fast but reasonably long-range 5G to 9.5 million people by 2024 on Sprint’s mid-band 2.5 GHz frequencies.
A lengthy and much-redacted filing with the Federal Communications Commission breaks down this pitch, starting with a few reasonably enticing bullet points:
- “an average download speed in excess of 100 Mbps or higher (with a minimum speed of 25/3 Mbps)”;
- a customer-installable router, with no extra monthly fee required for it;
- no setup charge;
- no contract.
That filing does not, however, renounce data caps. A mostly blacked-out paragraph on page 18 and a completely redacted chart on page 19 show that T-Mobile has calculated its own forecasts for customer use and compared them to reported monthly average household use at Altice (250GB in the fourth quarter of 2018) and median data usage at Comcast (174 GB in December 2018).
T-Mobile’s current smartphone plans enforce relatively generous limits on mobile hotspot usage, while the company’s $50-per-month invitation-only 50Mbps 4G LTE home service T-Mobile announced in March does not include data caps. The filing expresses confidence in the ability of Sprint’s 2.5GHz spectrum to add more capacity to the merged company’s network than its current business model will need in every market.
“When a 2.5GHz radio is deployed on a tower, it produces a very large increment to capacity all at once,” the filing reads. T-Mobile plans to combine those radios with the 600MHz spectrum underlying its own just-announced consumer 5G launch.
The filing continues:
While many of these areas need some 2.5GHz to be deployed in order to provide sufficient capacity to meet New T-Mobile’s performance thresholds, traffic in these areas would end up putting very little load on the 2.5 GHz spectrum, leaving much of it available to provide wireless fixed broadband service. The result is that New T-Mobile will gain substantial excess capacity in a number of areas for a very low cost.
The phrase “in a number of areas” does a lot of work there, warned New Street’s Chaplin.
“Sprint’s 160MHz of 2.5GHz would account for maybe a year or two of capacity growth in big dense markets,” Chaplin wrote of that carrier’s spectrum holdings. “It is certainly true that T-Mobile could pursue and even serve 10MM homes in certain markets. But that is only 8 percent of the country.”
A separate declaration to the FCC by T-Mobile President and Chief Operating Officer Mike Sievert offers a little more clarity. In it, Sievert—set to become T-Mobile’s CEO in May after the carrier’s infamously outspoken boss John Legere steps down—says the new 5G home-broadband service will cover 52 percent of US ZIP codes.
Sievert says that, by 2024, this new offering will reach deep into the territories of the two largest cable operators, touching 64 percent of Charter’s footprint and 68 percent of Comcast’s.
Sort-of-5G from some smaller rivals
Moffett expressed his own skepticism about using the scarcer spectrum of mid-band 5G for less-profitable fixed-wireless access. Smartphone users simply don’t chew through as much data as people sitting in front of laptops or connected TVs on residential broadband connections.
Instead, he pointed to a few experiments with what you could call “5G-esque” home broadband.
“There are other models out there that may ultimately prove more interesting than Verizon’s millimeter-wave approach,” he wrote. “Starry’s millimeter-wave-over-macrocells architecture has had some success in Boston, and mesh networks like Vivint and Common Networks on the West Coast have met with some success as well.”
Starry is the startup founded by Chet Kanojia after his prior venture—the broadcast-to streaming TV service Aereo—suffered death via a questionable Supreme Court interpretation of copyright law. It uses licensed 37GHz millimeter-wave spectrum for its $50/month, 200Mbps, no-data-cap home Internet service.
“If your fundamental question is ‘Will 5G allow you to dump Comcast?’ the answer is absolutely! Depending.”
Starry doesn’t have to hang a mobile service off its cell sites and so can optimize its network for fixed-wireless service.
In a 2018 research note, Moffett wrote:
Starry’s model for fixed wireless broadband is distinct from that of most of its competitors (most notably, Verizon) in that their service is delivered from macrocells rather than neighborhood-based small cells. Starry’s secret sauce is a proprietary hybrid analog/digital beam forming technology that, by ‘focusing’ radio signals, allows Starry to achieve much greater propagation distances than would otherwise be the case.
But of course, there’s still a catch. In Starry’s case, it’s that the company has, so far, confined its service to multiple-dwelling buildings in its initial markets of Boston, Denver, Los Angeles, New York, and Washington, DC. The company has not yet shipped a single-family home system.
Vivint and Common Networks, meanwhile, use mesh-network setups to distribute wireless broadband across neighborhoods in Salt Lake City and the Bay Area, respectively. If you don’t live in those places, that may not come as much comfort.
The last, best hope of broadband?
But for all of the engineering and economic hangups possible with 5G home broadband, at least this family of technologies has multiple companies working to make it happen in new markets.
You can’t say that for incumbent technologies like cable, where the last serious attempt at pushing one cable company to “overbuild” into another’s territory ended when the FCC lifted a condition governing Charter’s 2016 purchase of Time Warner Cable, or DSL, which is sandbagged by aging, un-upgraded infrastructure.
Google Fiber was supposed to crack open the resulting domination of many local markets by cable with breakthrough speeds and prices. But its rollout hasn’t just stalled—it’s reversed, with Google pulling out of Louisville this spring.
Smaller ISPs continue to build out fiber in local markets around the country, while low-Earth-orbit satellite constellations like those now being launched for OneWeb and SpaceX also bear watching. But wireless has the stronger theoretical potential to cover more ground more rapidly. And right now, 5G or something like it represents the best possibility for wireless to carry large quantities of data.
And while wireless has let us down before, 5G’s raw speed means it ought to work exceptionally well in some edge cases. Chaplin noted one that cable operators in particular might appreciate: extending service the last few hundred feet to a potential customer’s residence. Here’s hoping.