Your browser is not supported. Please update it.

17 May 2019

Special Report: The USA’s 5G race

The USA’s determination to ‘win the 5G race’ is leading to all kinds of unusual government interventions in the supposedly free market economy’s mobile industry.

The acquisition of Qualcomm by Broadcom was blocked, even though the latter was on the point of transferring its domicile to the USA from Asia.

Proposals to accelerate US 5G build-outs with a government-backed wholesale network are still doing the rounds.

5G is central to the escalating USA-China trade and security wars, which have led to the barring of Huawei and ZTE from infrastructure contracts, and may spark further sanctions.

And the Department of Justice has effectively warned a Federal Trade Commission judge to tread carefully before she issues any penalties against Qualcomm which might affect 5G progress.

Beyond issues of national pride and populism, it is not clear how important it is for either the USA or China to ‘win’ at 5G, or what that really means. The launches of fixed 5G by Verizon and AT&T meant the USA was ahead of China in providing commercial services, but those initial offerings are limited in scope and unavailable to the vast majority of citizens. China, given its superior levels of investment in building out enabling infrastructure, and its better mid-band spectrum position, is likely to get to wide area scale ahead of its rival.

But in the end, it is more important for commercial operators to have the right timing, and network coverage and capacity, and to align these targets with their business models and the key areas of demand. Being first with a network that cannot support popular applications, or doesn’t generate profitable ARPU, may be a feather in a government’s cap, but does nothing to deliver 5G’s promised social and economic benefits, or to improve the operators’ fortunes.

As more MNOs round the world start to launch 5G and to report on progress against meaningful KPIs, there will be a more realistic and nuanced picture of what is working, and not working, for 5G. The news that South Korea signed up 260,000 subscribers in the first month of turning on live 5G networks is a real indicator of a country that has invested heavily in new platforms, and which has the right market conditions to turn these into profit and new behaviors.

AT&T and Verizon expand 5G ambitions in case of a Sprint/TMO merger

The limited scope of the USA’s initial 5G launches, providing fixed wireless access (FWA) in a tiny percentage of the country’s landmass, bely the scale of the carriers’ ambitions to use 5G not just to enhance mobile services, but to achieve fully national broadband coverage. AT&T and Verizon aim to fill in the gaps in their wireline territories and reduce overall cost by combining fiber and 5G, sometimes to support converged fixed/mobile use cases, sometimes in a complementary way, using wireless to extend coverage. And T-Mobile has argued that, if it is allowed to acquire Sprint, it will be better positioned to roll out 5G nationally, and to support full broadband and multiplay services, including video, on it.

Randall Stephenson, AT&T’s CEO, told an investor relations event that his company expects to offer a “true, high speed Internet network throughout the United States” within 3-5 years using a combination of fiber and 5G.

“In 3-5 years out, there will be a cross-over point,” he said. “5G will cross over, performance-wise, with what you’re seeing in home broadband. We’re seeing it in business now over our millimeter wave spectrum. And there will be a place – it may be in five years, I think it could be as early as three – when 5G begins to actually have a cross-over point in terms of performance with fiber. 5G can become the deployment mechanism for a lot of the broadband that we’re trying to hit today with fiber.”

AT&T’s primary focus in this 5G and converged build-out is the enterprise. It does believe the fiber/5G combination will increase its share of the home broadband market – it claims to have an average of 25% market share in areas where it has deployed fiber, and says this could increase to 50% with the added value and stickiness of a 5G-based mobile service.

However, many of the operator’s most creative initiatives, such as its advanced network-on-demand offerings, are heavily focused on increasing enterprise revenues, and this is where the 5G build-out will mainly be justified, especially as future upgrades start to support more than just high speeds, and introduce low latency, high availability and massive device density.

“This changes how you think about utility management, pipeline management, traffic management. It produces all kinds of new business opportunities into our world, more than just speed,” Stephenson told the investors. “So the business side is where we’re spending most of our time. We’re going to be surprised about all the use cases that come from this.”

Stephenson also said AT&T will finish its government-mandated fiber build-out to 14m new locations next month – this was one of the conditions of the telco’s acquisition of DirecTV. However, that will not be the end of fiber expansion. “We’ll be deploying fiber until the day I retire,” the CEO said. “5G requires extensive fiber deployment.”

Stephenson’s vision of nationwide, high quality broadband is an ambitious one, given that he told the same event that AT&T would reduce its capex spend in the coming years. Like the other US operators, AT&T is disadvantaged in terms of 5G spectrum compared to peers in other advanced mobile economies. In most countries, launches are kicking off in the 3.5 GHz band, which supports high capacity without the engineering complexities and severe range limitations of millimeter wave bands. In the USA, 3.5 GHz is federal spectrum, which is being opened up to operators via a three-tiered system, but even when auctions are held this year for the licensed portion, operators will not secure the high bandwidth and virgin airwaves that MNOs in other markets are enjoying.

The big two US operators have led off with mmWave (28 GHz and 39 GHz) but while these support FWA services when operated at high power, for full mobile coverage, including indoors, they will require huge numbers of base stations, as well as US-specific devices. All this will add to the expense of hitting that “nationwide” target, even if “nationwide” turns out to mean urban and suburban rather than rural (like AT&T’s mobile 5G roadmap for 2020, which claims to be national but will actually reach 200m Americans in urban areas).

Many of Stephenson’s cost reduction goals relate to AT&T’s aggressive program of virtualization of its networks. The CEO said the company is on track to hit its target of virtualizing 75% of its network functions by 2020. “I’m actually quite optimistic that over the next couple of years capital spending actually works its way down,” he said.

There are more physical ways to save capex and operating costs – he described a “one-touch” project which will see contractors installing antennas for several bands in just one tower climb, as AT&T expands coverage in its AWS-3 and WCS spectrum as well as its 5G bands and the 700 MHz FirstNet public safety roll-out.

Verizon has been more circumspect about setting coverage targets while it is mainly using mmWave spectrum for 5G, and has confined its initial roll-outs to densely populated areas or FWA services.

Both the carriers will be aware of the fact that, if the Sprint/TMO merger is approved (see separate item), their spectrum disadvantage will be more glaring. TMO plans to roll out wide area 5G in its 600 MHz spectrum, which is very good for affordable coverage but will support very limited capacity (so the user experience will be little different from that in 4G). But if that is combined with Sprint’s 120 MHz of nationwide spectrum in 2.5 GHz, the combined operator will have an enviable balance of capacity and coverage, without having to take on the challenges of mmWave in the first phase of deployment.

If that merger happens, the big two telcos’ virtualization programs will be even more important to give them the agility, cost-effectiveness and service platform to respond to the threat of a truly national 5G competitor.

Verizon said last week that it had reached a key milestone, demonstrating a fully virtualized RAN baseband in its network in California, working with Intel and Nokia.

“It enables MEC [multi-access edge computing], and the ability to offer or host applications closer to the edge,” said Bill Stone, VP of technology planning and development at Verizon.

An earlier trial, held in January in Houston, Texas, used MEC equipment and software in a 5G testbed and claimed to have cut latency in half.

“Now that we can separate the software from the underlying hardware, it gives us more flexibility, and it’s important because it reduces latency. We also believe over time this is going to become a more efficient and cost-effective configuration,” Stone added. “The two things combined, California and Houston, are where we plan to go in the future. As we move forward with 5G deployment, it’s a critical part of our strategy to deploy these general purpose platforms toward the edge and take advantage of virtualizing more of the baseband functions while also preparing to host additional applications.”

The trial used Nokia’s AirScale All-in-Cloud Base Station architecture, Intel Xeon Scalable processor-based servers and the Intel FlexRAN reference architecture, which supports the RAN stack running on servers.

Verizon is working with other vendors including Ericsson and Samsung, and will conduct more vRAN trials this year. The telco is already progressing with virtualizing its core and plans to move towards 5G Standalone technology, with a cloud-native 5G core, next year.

“As we continue to proceed with 5G and modernizing the network, all of the new network functions that are being put in place will be cloud-native and fully virtualized,” Stone said. “This is table stakes.”

Adam Koeppe, SVP of network planning at Verizon, said in a statement: “Massive scale IoT solutions, more robust consumer devices and solutions, AR/VR, remote healthcare, autonomous robotics in manufacturing environments, and ubiquitous smart city solutions are only some of the ways we will be able to deliver the promise of the digital world. These advancements in technology are critical steps towards that realization.”

Sprint gets desperate for TMO as network failings weigh on results

Sprint used to have the reputation for being the most technology-driven of the major US operators when it came to its network. It tended to be the first to try a new technology or experiment with a different approach, in the quest to differentiate itself on the quality of its infrastructure.

But two major missteps turned its network ambition into an albatross round its neck. One was the adoption of WiMAX as its initial 4G platform, in a bid to gain a headstart on rivals – in fact it was left having to migrate away from a stranded technology and ended up years behind the LTE leaders. The other was the botched acquisition of Nextel, which should have brought Sprint a loyal customer base and some high value service offerings, notably push-to-talk. Instead, the integration of Nextel’s IDEN network (another technology island) was mishandled and had disastrous effects on the quality of experience for those subscribers, who left in droves.

Sprint has not yet fully recovered from those mistakes, or from the failure to achieve its technology goals, or to turn its 4G build-out into commercial success. Its ambitious Network Vision program to move to a unified platform supporting multiple radios and many spectrum bands – and other operators on a hosted basis – overran in time and budget and did not deliver the smooth user experience customers wanted. It was replaced by another initiative, but the results have been slow to become clear.

None of this bodes well for its 5G future, especially if regulators prevent its proposed merger with T-Mobile USA. Sprint brings its only unique asset to that marriage – its plentiful supply of 2.5 GHz spectrum, the kind of midband airwaves that are ideal for 5G capacity, without the risks of millimeter wave, and which the larger operators lack. It would complement TMO’s national 600 MHz spectrum very well, supporting a balance of coverage and capacity.

Most of the other benefits of the merger come from the other side, particularly TMO’s brand and its strong appeal to consumers. While TMO has used marketing expertise and canny pricing to leapfrog its would-be merger partner, Sprint has suffered years of customer defections, caused by network failings and poor customer service. Despite big investments in these areas in recent years, and a major change of management, its fiscal fourth quarter results were depressingly familiar, with 189,000 postpaid customers leaving – a sad reversal of the year-ago quarter, when Sprint reported net additions of 55,000.

CEO Michel Combes said in his statement: “While we’ve made progress, there are certainly continued challenges to address, which will continue to put pressure on our service revenue and retail customer growth.” Those challenges resulted in a net loss of $1.9bn for the operator’s full fiscal year, reversing net income of $7.4bn in fiscal 2017 (though the contrast largely related to one-off charges and benefits).

But the company still needs to invest more in network quality and customer experience, in its seemingly never-ending push towards sustainable growth in subscriber numbers and profits.

In its earnings statement, Sprint said: “While the company continues to look for opportunities to improve operational and cost efficiencies in fiscal 2019, these improvements are expected to be fully offset by incremental costs associated with network and customer experience initiatives.”

All this casts doubt on how Sprint can survive if it is blocked from merging with TMO. It would almost certainly try to seek another white knight, perhaps a large cableco, or even its former suitor Dish Network (which was gazumped by Softbank) – though the latter would present many complications in terms of integration and business model.

Combes said Sprint had recently filed a document with the FCC and Department of Justice to explain the challenges that Sprint faces and what 5G deployments would look like with and without the merger. He added: “We still lack low band spectrum. We still lack scale, and we still lack financial resources” relative to AT&T and Verizon.

“If you look at the history of Sprint, the network has suffered from years of under-investment,” CTO John Saw told FierceWirelessTech at the Brooklyn 5G Summit. “For years I realized that I am not able to invest as much as my competitors, especially Verizon and AT&T”, and that has led to a smaller footprint and more reliance on roaming. That is worsened by the fact that Sprint has less than half the sub-GHz spectrum of its competitors – while its 2.5 GHz airwaves are good for capacity and dense urban deployments, they make wide area and rural roll-out very expensive because of their limited range.

Saw acknowledged that Sprint did not bid in the 600 MHz incentive auction because it could not afford to (TMO was the biggest buyer) and added: “We’re not saying it’s a complete end of the world for Sprint, but Sprint is going to be handicapped moving forward … When you couple the 2.5 GHz with T-Mobile’s 600 MHz, it’s basically a match made in heaven.”

The pleas that Sprint may fail without TMO, and that merger is the only way to enable effective competition to the big two, may still fall on stony ground. Many analysts believe the odds against merger approval are lengthening, and a client note from  MoffettNathanson said: “Given the mounting odds against merger approval, one would expect growing attention to Sprint’s results. The reality of the picture at Sprint—the subscriber base is shrinking, ‘real’ EBITDA is flat to declining, and free cashflow is barely positive.”

For all the business challenges inherent in Sprint, TMO still needs the $26bn deal too. Without it, it will be severely disadvantaged in terms of 5G spectrum and unable to support high bandwidth services, some of which, like its video plans, are central to its growth strategy.

So the companies are getting more desperate in their pleas to the regulators, and they are now reported to be considering new concessions including the possible separation and sale of their respective prepaid divisions. Sources told Bloomberg that other options include selling spectrum licences or supporting the creation of a fourth operator via network leasing deals (a popular condition in European Union mergers which would reduce the number of MNOs in a market).

However, the sale of one or all the prepaid units is reportedly the most favored concession as it would address a key concern in some states – the reduced competition in prepaid, which would potentially raise prices and reduce choice for low income citizens. Together, TMO’s Metro brand, acquired with flat-rate carrier MetroPCS, and Sprint’s Boost and Virgin Mobile, control 42% of the US prepaid market; followed by Tracfone on 32% and AT&T’s Cricket on 25%. Merger would mean only three meaningful sources of prepaid services.

The two companies have already promised to create a customer service center employing more than 1,000 people near Rochester, New York if the merger is approved.

Despite all these uncertainties, Sprint has to carry on with at least a semblance of business as usual and it claims it will launch commercial 5G services in four markets “in the coming weeks”. On its quarterly earnings call, it said it might have a larger initial 5G footprint than Verizon or AT&T did, focusing on Chicago, Atlanta, Dallas and Kansas City. Five more will follow by the end of June – Houston, Los Angeles, New York City, Phoenix and Washington DC, with the initial footprint across all nine cities expected to be more than 1,000 square miles, according to Saw.

Saw reiterated his plan to leverage 2.5 GHz spectrum and Massive MIMO to “simultaneously support both LTE and 5G, so it’s like killing two birds with one stone. I can see us moving forward to continue to invest in Massive MIMO.”

Sprint now has 2.5 GHz spectrum deployed on about 80% of its macro sites compared with 60% a year ago. It also has about 30,000 outdoor small cells deployed, including mini-macro and strand-mounted form factors, and it has deployed about 1,500 Massive MIMO antennas, Saw said. The new radios can be upgraded to mobile 5G via a software update.

Mishka Dehghan, VP of 5G development at Sprint Business, told last week’s Big 5G Event in Denver that Sprint’s Curiosity IoT platform, introduced last September, will enable some of the new 5G-based revenue streams, including smart city services, which are  being trialled in testbeds in Greenville, North Carolina and Peachtree Corners, Georgia.

The latter features an urban transport system that uses micropositioning for real time location tracking of objects, with accuracy down to an inch, compared to GPS’s 5-10 meters. This, with HD mapping, could enable autonomous vehicles.

Sprint also has a virtualization program, and plans to be the first US operator to launch VoLTE on a fully virtualized core this year.

“We’ve delivered a very good voice experience on an efficient wireless network platform and wanted to be sure of the quality of experience before transitioning to VoLTE,” said the company. Sprint currently uses CDMA for voice, so its subscribers cannot use voice and data at the same time. TMO already has VoLTE, which now accounts for 88% of its voice calls, though it has not deployed it on a virtualized core.


DoJ intervenes in FTC case against Qualcomm, to protect US 5G leadership

The US Department of Justice (DoJ) has taken the highly unusual step of, effectively, cautioning a federal judge against tough action in an antitrust case. The DoJ filed a

“statement of interest” concerning the Federal Trade Commission’s (FTC’s) suit against Qualcomm, warning Judge Lucy Koh that too negative a ruling could disrupt the USA’s 5G progress.

The DoJ told Koh – famous for presiding over several of the landmark patent battles between Apple and Samsung – that if she decides against Qualcomm in the San Jose trial, she should hold a hearing at which the company could outline the impact of penalties on the emerging 5G market.

Andrew Finch, a principal deputy assistant attorney general, wrote in the filing: “There is a plausible prospect that an overly broad remedy in this case could reduce competition and innovation in markets for 5G technology and downstream applications that rely on that technology. Such an outcome could exceed the appropriate scope of an equitable antitrust remedy. Moreover, it has the distinct potential to harm rather than help competition.”

The key aspects of the FTC case are:

  • Qualcomm’s licensing practices are not compliant with Frand (fair reasonable and non-discriminatory) terms, and that has harmed the cellular industry
  • Licensing based on the price of the whole device, not just the chip, is unfair
  • Qualcomm’s market power combined with its licensing policies have harmed competitors such as Intel

The FTC has asked the San Jose court to ban Qualcomm from requiring its chip customers to take a patent licence. If it succeeds, it would undermine a cornerstone of the vendor’s controversial but very successful business model, namely its ‘no licence, no chips’ policy. If it loses the case, it may have to renegotiate some of its licensing deals and fees, and to license standards-essential patents to rival chip suppliers under Frand (fair reasonable and non-discriminatory) terms.

But the DoJ called on Koh to refrain from setting penalties without further hearings, if she decides against the chip vendor. Finch wrote: “Because an overly broad remedy could result in reduced innovation, with the potential to harm American consumers, this Court should hold a hearing and order additional briefing to determine a proper remedy that protects competition while working minimal harm to public and private interests.”

Lawyers expressed surprise at the filing, and said it was extremely rare for the DoJ to interfere in FTC proceedings. However, the DoJ offered the precedent of an appeal court’s ruling in the US antitrust case against Microsoft, which said that the vendor had “the right to judicial resolution of disputed facts not just as to the liability phase, but also as to appropriate relief”.

There is a political angle of course – not just the government’s desire to protect any aspect of 5G leadership, but also, according to a Reuters report, the Trump administration’s wish to reverse cases brought under the previous presidency, and backed mainly by Democrat FTC commissioners.

“Remedies should not crush a tiger’s spirit; they should train, not tame. Among other things, this means that equitable remedies should not interfere with the defendant’s innovation incentives going forward,” Finch wrote, quoting one legal scholar.

Testimony in the case ended in late January and a decision is expected very soon. After the hearings had ended, Apple and Qualcomm settled their own string of lawsuits over licensing, in mid-April, leaving the FTC proceeding as the only active case against Qualcomm in the USA.

Some experts believe the settling of the disputes with Apple – which was a major instigator and witness in the FTC suit – has weakened the federal case significantly.

They argue that, by signing a long term licensing contract, Apple has effectively acknowledged that Qualcomm’s licensing practices are fair, and this is strengthened by the 100+ other licensing deals the chip company has lined up.

Disclosures made during the Apple trials tend to support Qualcomm’s arguments that its fees are fair, given the importance of its IPR – testimony said that Qualcomm licensing fees are up to 5% of the wholesale price of the phone, with a per-device price cap of $400. That gets licensees access to more than 130,000 patents.

Meanwhile, Qualcomm’s latest licensing deal is with HMD Global, the handset start-up which makes phones under the Nokia brand. HMD has signed a patent agreement which allows it to develop devices that use Qualcomm’s 3G and 4G single-mode and new 5G multimode chips. HMD’s Nokia-branded handsets have become the biggest sellers in the featurephone sector and the vendor is ranked thirteenth in smartphones, according to Counterpoint Research. Last year, HMD raised $100m to expand its business.

Wholesale 5G network still being pushed in the USA, despite FCC hostility

Last year, there was a flurry of discussion about the pros and cons of state-backed wholesale mobile networks, as a means of accelerating progress and opening up the platform to a wider variety of service providers.

This was prompted by comments from US president Donald Trump and later his re-election committee, apparently backing such a scheme in the name of the ‘5G race’. But this was not an off-the-cuff remark by the president, but an idea that has been discussed in various government departments for some time – and judging by recent comments from FCC Commissioner Michael O’Rielly, is still a live issue now.

O’Rielly is deeply opposed to the idea, as he articulated in a blog on the FCC website.

“Over the last few months, various ideas have been floated about the offering of 5G wireless services via a government-sponsored network,” he wrote. “This entire effort seems convoluted and borders on the preposterous.”

While there are plenty of reasons why a wholesale-only network can drive new competition and services, state-run approaches have rarely been a success. In fact, the USA has a potential private sector route to a shared 5G platform, if Ligado is allowed to pursue its plan to build out a wholesale network in its mobile satellite spectrum. And Dish’s own roadmap, to roll out first NB-IoT and later 5G, also envisages a wholesale business model supporting many service providers. The impact of these projects could be greatly intensified if Amazon AWS, or another webscale company, decides to partner with Dish or Ligado.

By contrast, a state-driven approach has not attracted much favor in the USA and even Trump distanced himself from the previous comments and insisted 5G would be driven by the private sector. And earlier this month, FCC chairman Ajit Pai told the National Spectrum Consortium’s 5G Collaboration event: “Like a recurring bad dream, there have been repeated calls for government control of America’s 5G networks. Let me be clear: I oppose any proposal for the government to build, own, or operate a commercial next generation wireless network.”

But it seems that some parties are still pushing for the state-run plan, since both Pai and O’Rielly are going to such pains to debunk it. Among the latter’s reasons to oppose such a plan were:

  • It has been “nearly impossible to nail down with any granularity” the details of how such a plan would operate
  • There is no available spectrum to be used for the network
  • The government would not be able to compel telcos to use the network, so it might find it hard to attract sufficient business
  • The government has no existing tower contracts, so the network would have to built from scratch
  • A government-led wholesale network would not be any more secure than a private one, and might be more exposed to attack

There have always been technical and commercial obstacles to shared networks ideas, whether state or private. Regulators are concerned about diminished competition if active infrastructure or spectrum are shared (ignoring the fact that the most innovative services are coming from MVNOs or over-the-top providers). MNOs tend to guard control of their network jealously, seeking differentiation via the best sites and network optimization. There are standards for multi-operator RANs, including small cells (where sharing is most essential), but specs like 3GPP MOCN are underused.

This explains why some governments have sought to kickstart new network build-out while reducing the MNOs’ cost burden, by creating a wholesale open access network under the auspices of a state initiative, usually with some investment from private backers and possibly the MNOs themselves.

This is an approach which the operators’ representative body, the GSMA, opposed fervently in a report published last year, ‘Wholesale Open Access Networks’.

It states at the beginning of the document: “Some supporters claim that these networks will deliver greater coverage than market competition can. However, those making this claim often gloss over the fact that, in order to be built, the SWN [single wholesale network] or WOAN [wholesale open access network] require significant public subsidies and other forms of support which are typically not available to competing network operators. The GSMA believes that network competition can and does deliver mobile network coverage. In areas where building networks is uneconomic there are other approaches. They include voluntary network sharing that can facilitate coverage in a particular area.”

But shared resources will be essential to 5G success, to extend coverage deep indoors and to remote areas which do not support conventional MNO business cases well, but are essential for ubiquitous IoT services.

The failure of leaving everything up to the MNOs themselves is seen in attempts by various countries to push the operators into voluntary sharing arrangements (sometimes with the threat of compulsion) in rural areas. From the UK to Australia, these efforts have resulted in disputes and litigation, but not satisfactory ubiquitous services, though the UK is currently revisiting the issue (see separate item).

If well thought-out, a wholesale network with external or state investment could improve the business case for MNOs, at least partly addressing their complaint that, in a market-based system, they cannot justify investing in unprofitable networks such as those in remote areas. Yes, open access brings new competitors, but those exist anyway in the unlicensed spectrum world, and companies like Google and Federated Wireless will enable a whole new approach to wholesale access, operating geolocation databases and, in future, on-demand spectrum services, in shared or unlicensed bands. And MNOs still have advantages over any new rivals, with their extensive customer reach, brand awareness and apps partnerships.

But these arguments appear to carry little weight with the GSMA. Its report examines five SWN or WOAN projects, in Kenya, Mexico, Russia, Rwanda and South Africa, in a follow-up to its 2014 study by Frontier Economics, which assessed the economic case for a wholesale network model. It makes valid points about the delays and political disputes which have dogged progress in nearly all state-inspired WOAN initiatives. But it takes a very rosy view of MNOs’ progress in voluntary infrastructure sharing, which almost always stops short of active RAN pooling, and claims its members are “also exploring new business models with third parties to share the cost and risk of investment in rural and remote locations”.

There have certainly been many problems and delays in the wholesale projects cited in the report, and these have been covered in the past in Wireless Watch. Russia’s supposed neutral host operator, Yota, ended up being acquired by MNO Megafon and the operators went their own way on 4G (though the government’s shared network plan certainly pushed them into action). And Kenya’s project seems to have been shelved amid MNO opposition.

Two other projects are criticized mainly for being delayed, meaning that there are no real results to be assessed. Mexico has finally awarded the 700 MHz spectrum, and contract to build and run its wholesale LTE network, to a consortium called Altran and this will be eagerly watched in a country which is trying to reduce the power of the incumbent Telmex and introduce new competition and services.

Meanwhile, South Africa has had several false starts, but in October 2017 published an ambitious plan which would, if accepted, up-end the country’s approach to spectrum licensing and access. It proposes a public-private partnership (PPP) to develop an open access wireless network, with the consortium including a wide range of participants including the MNOs, private equity, OTT and ISP players and infrastructure companies. It proposes to use all the spectrum available rather than assigning just one band. This is a radical approach, and will come with risks and a significant dose of political conflict. However, it does set out a vision which will, we believe, become far more normal during the second phase of 5G, when extreme densification, network slicing and vertical/IoT services transform the mobile model for everyone.

By contrast with these as-yet untested projects, Rwanda has a live WOAN, which launched on schedule in late 2014 in capital Kigali as a PPP between the government and Korea’s KT, called olleh Rwanda networks (oRn).