The hype surrounding the USA’s 5G program died a little over the past week, with the prolonged 28 GHz spectrum auction attracting limited interest, and AT&T being castigated (again) for marketing LTE services as ‘5G’. However, behind the obsessive focus on the USA’s first, small-scale roll-outs, and its status as the home of the ‘world’s first’ mainstream 5G services, there are genuinely interesting plans afoot. This is largely thanks to the fact that AT&T and Verizon are actually doing what most operators have only spoken about, and tying 5G to broader new platforms such as virtualization and on-demand enterprise services.
AT&T has set out its roadmap for 5G services, which sounds more concrete and real world than most operators’ ‘visions’. This is for two reasons. One, 5G is a central part of the company’s broad ‘Domain 2.0’ program to introduce virtualization and software-defined networking across its networks, fixed and mobile, transforming its economics and its supply chain. By contrast, though most operators talk about 5G and virtualization being complementary, most are rolling out early radio upgrades well before their digital platforms are mature or even planned.
Two, AT&T is focusing on enterprise services. Almost all operators agree these will be the main way to drive new revenues from 5G, and therefore achieve decent ROI. But the majority of MNOs are deploying 5G initially to support faster data rates for established use cases, most of them geared to consumer mobile broadband or connected cars. There is, therefore, a big gap between the first wave of deployment and capex spending, and the concrete plans to support those new use cases and revenue streams in the enterprise and IoT.
AT&T has set out “three pillars” of 5G strategy, which it says will add up to a toolbox to help enterprises adopt 5G-enabled applications to help transform their own businesses. These pillars are fixed wireless – the first focus of its current roll-out; edge computing; and a rapid move towards nationwide coverage. That combination will support the full range of enterprise requirements, including ubiquitous coverage for applications like asset tracking; targeted fixed capacity for very high bandwidth use cases in specific locations, such as VR-assisted training; plus localized cloud computing resources to add data analysis to the enhanced connectivity of 5G, for instance, to provide rapid AI-enabled decision support for humans or robots, on-premise.
As it expands 5G coverage, gaps will be filled by continuing to enhance LTE, including for fixed wireless. “In the coming weeks, we will offer multiple speed tiers up to 50Mbps” based on LTE-Advanced, it said. While it now has fixed 5G available in 12 cities, with the Netgear Nighthawk device, LTE-A is being rolled out to 400 markets, leveraging AT&T’s fiber network, which the operator says now comes within 1,000 feet of 8m business customer locations in the USA, and connects 2.2m sites.
Full mobility will be the next step. Although, like Verizon, AT&T has deployed its first 5G networks to support fixed wireless access (no handsets required), it has promised to add mobile services rapidly, even in its 39 GHz spectrum. However, true wide area mobile services will require lower frequencies – the cost to achieve broad coverage with the limited range of millimeter wave (mmWave) spectrum would be astronomical ($350bn to $570bn for the whole of Europe, says the head of Deutsche Telekom, so the USA would be comparable in scale).
AT&T has not yet said which lower band spectrum it will use to complement its mmWave high capacity roll-outs, but says it will start these deployments in the second half of 2019, and will have its first smartphone ready then, sourced from Samsung, and able to support both mmWave and sub-6 GHz frequencies.
The operator has been deploying ‘5G-ready’ radios in several bands, with a view to migrating all of them, over time, to 5G as older services are switched off. Its lowest frequency assets are in 700 MHz, since it was not a big buyer in the 2017 600 MHz incentive auction (and is reported to have sold off the licenses it did win, once it secured additional 700 MHz airwaves thanks to its FirstNet public safety contract). It is 5G-migratable radios for FirstNet but its other 700 MHz assets are less likely to support near term 5G. They are still needed for LTE coverage (which will remain central to most MNOs’ services for many years) and are based on older radios with no 5G readiness. When the time comes, then, they will need to be entirely replaced to support 5G.
Other sub-6 GHz options where AT&T says it has 5G readiness, as well as spare spectrum, include AWS-3 (1695-1710 MHz/1755-1780 MHz/2155-2180 MHz) and WCS (2.3 GHz).
AT&T claims to cover 500,000 square miles with the FirstNet network as of the start of this year, which would give it an extensive, modern grid on which to activate 5G and gain a significant headstart in population coverage terms – with some of the costs allayed by FirstNet contract revenues. This could put the operator in a strong competitive position against the other MNO which is focused on broad coverage in the near term, T-Mobile. TMO was the biggest buyer of 600 MHz licences, which will support cost-effective wide coverage, but it will be building its network from scratch, with no existing revenue streams.
Verizon and Sprint are taking different approaches, at least during 2019, focusing on targeted capacity. Verizon will stick to FWA in its 28 GHz roll-out, for longer than AT&T it seems, though it claims to have enough spectrum across mmWave and sub-6 GHz bands to support its 5G needs for the foreseeable future. Sprint is heavily focused on its midband 2.5 GHz spectrum for a 5G capacity play – which, if it is allowed to merge with TMO, will complement that MNO’s strategy. Although 2.5 GHz is not a prominent global 5G band, having been allocated mainly for LTE, it is an official 3GPP designation and Sprint has carried out 5G New Radio tests and trials with, it says, very strong results in spectrum which provides a good balance between capacity and coverage.
Despite AT&T’s apparent advantage, some analysts believe TMO will catch up rapidly. Walter Piecyk of BTIG Research thinks the ‘Uncarrier’ will win the race, writing in a client note: “We believe T-Mobile is best positioned to deliver real 5G … T-Mobile’s deep and unused low band spectrum should enable it to launch real 5G before all of its peers.” TMO spent $8bn on 30 MHz of 600 MHz spectrum nationwide.
However, if the Sprint merger is not approved, TMO is likely to have to spend at least $20bn to acquire enough capacity-focused spectrum to complement its low band assets.
The operator has hired Ulf Ewaldsson – a 30-year veteran of Ericsson – as SVP of technology transformation to drive the 5G technology evolution, reporting directly to CTO Neville Ray. Most recently, Ewaldsson was chief strategy and technology officer, head of group function strategy and technology, and a special adviser to the CEO. However, he was closely associated with former CEO Hans Vestberg, so his departure was expected once new CEO Borje Ekholm started to wield his new broom. Vestberg has also turned up at a US operator, as CEO of Verizon.