Some 12 years ago, EE was given a controversial advantage in the 4G race when Ofcom, the UK regulator, allowed the company to run with previously licensed 1800MHz spectrum. The other three networks, to their horror, had to wait for an auction of new spectrum licenses before they could move off the 3G track and start in 4G. Until that auction took place, it was like watching an athlete bounce along in those Nike Vaporfly trainers, with miniature trampolines for soles, while their rivals clatter around in clogs. Something similar could now happen in 5G.

All four UK networks launched 5G years ago, but not all of them provide the latest version of the technology. The older one, called "non-standalone" 5G, bolted a new radio access network (RAN) onto the old 4G core. Its "standalone" successor introduces a 5G core to pair with the 5G RAN. Promising various benefits and features, it has been described before now as "real" or "true" 5G by its supporters, critical of the telecom industry for not prioritizing the new core at the very outset. Across most of the UK, not to mention numerous other markets, standalone remains widely unavailable.

Until last week, only two networks supported a 5G standalone service. Vodafone became the first telco to launch it in June last year, marketing it under the brand of 5G Ultra. But only the cities of London, Manchester, Glasgow and Cardiff were named as places where standalone was available. This was the same month in which Vodafone also announced plans to merge with Three. It has positioned this deal as its only hope for survival in a market where it cannot afford to compete. The shortcomings of the standalone network are offered as proof.

Promoting the bad

Ever since, then, Vodafone has turned normal marketing practice upside down, flaunting its deficiencies rather than its strong points. "There are only two players with the scale to invest to a level the UK deserves," said Andrea Donà, Vodafone UK's chief networks officer, in April this year. He was not, of course, speaking about Vodafone and Three. His company's 5G standalone network was unavailable outside "pockets" of the country, he freely admitted.

Virgin Media O2 (VMO2), the other telco providing standalone 5G before last week, and the one Donà had in mind besides BT, claimed to have switched on a service across 14 cities in February this year. VMO2 now boasts a service in as many as 300 towns, implying it is very widely available. Yet the company has not disclosed estimates of population coverage and would not share them when asked by Light Reading. It also ranked as the worst of the four networks for outdoor 5G population coverage, with a score of 51%, when Ofcom crunched the numbers for January this year.

That leaves Three and BT. The former, which appears to have not so much as dabbled in standalone, seems even keener than Vodafone is to portray itself as a network loser that only a merger (with another loser) will save. In the absence of that merger, which UK competition authorities are still reviewing, Three's message is that it can barely afford to keep the lights on. "Our cashflows have been negative since 2020 and our costs have almost doubled in five years, meaning investment in network is unsustainable," said Robert Finnegan, Three UK's CEO, in canned remarks accompanying its most recent set of results.

The full spectrum

All this left BT, which bought the EE network almost nine years ago, in a potentially advantageous position when it launched what has been advertised as a widely available 5G standalone service last week. According to the operator, it already blankets 18 million people across 15 of the UK's biggest cities and towns, where BT's 5G network covers at least 95% and usually 98% of the population. "It was important for us to make sure that we had that really reliable outdoor coverage because we felt not everyone does that," said Howard Watson, BT's chief security and networks officer.

Unlike in 4G, the lead BT clearly enjoys over rivals is less obviously to do with regulatory largesse. Even so, BT does have access to far more spectrum than any of its rivals – some 360MHz altogether, according to data included in a recent report from the Competition and Markets Authority, compared with Vodafone's 267MHz, Three's 250MHz and VMO2's 242MHz. By combining spectrum across multiple bands, a technique dubbed "carrier aggregation," BT reckons it can offer a higher-speed service on standalone than it could on non-standalone 5G.

It is not just about the amount of spectrum, either. Three's licenses are concentrated in the 3.4GHz to 3.8GHz range, good for speed, bad for coverage (especially indoors). Even after a merger, Three and Vodafone would have less spectrum between 1400MHz and 2.6GHz, holding 202MHz in total, than BT does with its 230MHz allocation across multiple bands. Using some of that in conjunction with its wall-penetrating 700MHz frequencies will boost indoor availability, according to Watson. With non-standalone 5G, he explains, BT has been hampered by its reliance on 1800MHz spectrum, poorer for coverage, to connect to the 4G core.

Progress on ripping out Huawei and extending 5G RAN coverage with Ericsson and Nokia has undoubtedly helped. Older Huawei kit is unlikely to support 5G standalone and the Chinese vendor's 5G products must, in any case, be replaced by the end of 2027 under government rules. But Huawei does not figure in the 15 cities and towns where BT has launched standalone. Upgrades to Ericsson and Nokia basestations have been carried out across some 30,000 cells spanning these locations.

Ofcom's data seems to provide evidence of the coverage gap between BT and its three network competitors, one that could prove significant in standalone. In January, the regulatory body put BT on 74% for outdoor 5G population coverage, with Three on 67%, Vodafone on 57% and VMO2, as previously noted, trailing on just 51%. A customer losing 5G standalone connectivity, and falling back to 4G, is likelier to notice the difference than one dropping off non-standalone. For more advanced standalone services, the downward gear shift could be disruptive.

Their merger case has left Vodafone and Three in an awkward position. Pleading poverty, Three has slashed capital expenditure from £334 million (US$437 million) in the first half of 2022 to just £230 million ($301 million) in the same period this year. Vodafone's argument is also that it cannot justify lavish expenditure given such poor returns on investment in today's four-player market. As they wait for a regulatory decision on a merger, they face an apparent choice of falling even further behind or giving satisfaction to the critics who insist rivalry spurs investment.

Rotting cores

What's more, the status of the 5G RAN is not the only issue. BT this year completed its move from an old 4G core provided by Huawei to a new Ericsson converged core, supporting 4G, 5G non-standalone and 5G standalone services. Thanks to parallel investment in a telco cloud, built with the help of UK software company Canonical, it has been able to spread the control plane, the network decision-maker, across eight sites and distribute the traffic cop of the user plane across 16. This rearchitecting allows BT to move workloads around the network, and the compute-intensive nature of artificial intelligence (AI) will make that more important in the future, according to Watson.

By contrast, Three's own mobile data core was looking outdated at the start of the year. Provided by Nokia, it was already five or six years old and not based on the latest container technologies, complained Anil Darji, Three's chief network architect, at Informa's Network X event in October 2023. Alternative vendors were being considered as part of a tender, said Three. But it had not made progress when Light Reading caught up with David Hennessy, Three's chief technology officer, in January, and the operator has provided no update since then.

As for Vodafone, it has also been shifting to another vendor, moving from a Cisco 4G core to a converged Ericson core that seems to resemble BT's. But it has been hush-hush on progress since early 2023, when the Cisco platform was still in use and executives declined to say when it would be scrapped. Ditching it would remove what Donà called "regrettable spend" at the time, but the migration of customers from one core to the other involved "lots of heavy lifting," he said. That is partly because the old core was installed at just a few sites. The goal with its successor was to grow that number.

A jumble of vendors

None of this, of course, means BT has conclusively won the standalone battle. A Vodafone-Three merger without stringent remedies would be instantly transformative, giving the new company about 26,000 mobile sites (after disposals), compared with BT's 19,000, and 517MHz of spectrum, including 59% of all licensed frequencies between 3.4GHz and 3.8GHz (although Vodafone has offered to sell a chunk of spectrum to VMO2 if the deal is allowed). It promises to spend £11 billion ($14.4 billion) over ten years on blanketing the UK with a 5G standalone service.

But it would have a complicated jumble of vendors across RAN and core. Three used Nokia for 3G and Samsung for 4G before naming Huawei its sole 5G supplier several years ago. It introduced Ericsson when the government ordered telcos to remove Huawei's 5G products by the end of 2027, but equipment from all four vendors is still present in Three's network. Samsung remained active at 7,000 sites at the start of 2024. Some cities, including Glasgow, host every one of those suppliers. And Three was adding Mavenir to the mix as a "small cell" provider earlier this year.

Vodafone, meanwhile, has shifted or is moving to Ericsson for all but 2,500 of its roughly 18,000 sites. Huawei is the legacy provider at the remainder, but Vodafone is replacing it with Samsung and NEC, plus IT components and software from vendors including Dell, Intel and Wind River. This "open RAN" deployment, however, has been remarkably slow. After notifying the world of its plans in November 2020, Vodafone took about two-and-a-half years to choose vendors and trial products before it kicked off a commercial rollout in August last year. In April, it declined to provide a status update.

The merging players would also have multiple core networks, with Ericsson, Nokia and possibly Cisco all in use. Maintaining the lot would add to the cost and complexity of the operation. Given Three's concerns about the mustiness of the Nokia technology, Ericsson seems the likeliest winner if the operators decide to move to a single core. Indeed, Three may prefer to wait for a regulatory decision on the merger than press on with its tender. But the experiences of BT and Vodafone indicate that migrating subscribers between cores is not easy.

Perhaps the main risk for BT is that customers simply do not care about standalone or resist the higher prices it is attaching to standalone gadgets and plans. There is a hint of "build it and they will come" mentality in BT's approach, with the operator girding itself for a torrent of AI applications that might not appear. Even if they do, they might not require standalone networks as a foundation. And while VMO2's standalone network seems less widely deployed, it is not charging a premium. Nor, for that matter, is Vodafone.

"EE failing to cash in on 4G head start, says City," reads an Evening Standard newspaper headline from February 2013. That was just a few months after the launch of 4G, when bank analysts were disappointed the new generation had not boosted EE's revenues. After the experiences of the last decade, few will be optimistic that 5G standalone can make that all-important difference.