Midway through Donald Trump's presidency, the companies that operate the UK's mobile networks would have realized they were staring at a potential multi-billion-dollar problem. Like other European operators, they had spent the last decade bingeing on Huawei. Lauded for products both low cost and advanced, the Chinese kit vendor had displaced Ericsson and Nokia, the Nordic old timers, as the mobile supplier of choice. Suddenly, America's president was railing against Huawei as an unacceptable threat to security. Courting Trump's favor after Brexit, the British government looked willing to intervene.
It eventually ordered telcos to evict Huawei from 5G by the end of 2027, forcing them to rip out basestations and install new equipment at significant expense. Worse, after many years of consolidation, that ban has left the market with only Ericsson and Nokia as realistic alternatives, say the operators. Each vendor has recently experienced its share of difficulties, ranging from periods of unprofitability to charges of corruption and resulting fines. If one became unviable for any reason, an equipment monopoly would exist.
A desire to avoid this scenario is the main reason for the interest in open RAN, a senior telco executive told Light Reading. For all the attention it has received, open RAN is basically just a set of interfaces that will supposedly allow one vendor's products to work readily with another's. These were previously unavailable, and so operators have tended to buy everything for a given site but especially the radios and baseband (signal-processing) gear from the same vendor's system. With open RAN, operators should be able to mix suppliers, and suppliers should be able to specialize. A radio maker would not require a baseband product to be competitive.
The hope remains that open RAN will breed competition and give operators the choice they crave. Yet there is growing industry doubt that supplier diversity is either achievable or sustainable. One possibility is a changing of the guard, as today's giants are replaced by a relatively small number of open RAN successors. Another is that Ericsson, Nokia and even Huawei capture a big share of open RAN business, leaving minimal room for others. Diversity looks feasible at the fringes in the market, say, for campus networks. But few expect a galaxy of choice for macro deployments.
That is partly because the main competitive premise is fundamentally flawed. Open RAN, it is true, could aid specialization. If it eventually allows vendors to be combined without undue hassle, a company like Mavenir could thrive as a standalone provider of baseband software. Any operator buying products from standalone providers would naturally use more suppliers than one purchasing the old-fashioned way. But Mavenir's baseband software is not a substitute for a radio made by Ericsson or KMW. Both are needed.
For open RAN to bring genuine diversity, there would need to be lots of Mavenirs, plus lots of KMWs. In fact, there would have to be lots of competitors in each part of the open RAN supply chain, from chips all the way through to software. With too few suppliers in any field, operators would risk the same dependency they have on Nordic suppliers now. The weaker the supplier, the more dangerous it could be.
Yet there is no rational explanation why each of these sub-sectors would necessarily feature more players than today's broader market does. Far likelier, judging by experience, is that consolidation will shrink each of these pools to a few economically viable companies. At the turn of the century, the mobile equipment industry featured numerous vendors that either collapsed or were swallowed up in mergers as mobile went mainstream. Alcatel, Lucent, Marconi, Motorola and Nortel are just a few that no longer exist in their historical form.
While nobody wants an equipment monopoly (except perhaps the monopoly), overcrowded markets can be just as undesirable. Smaller players simply cannot realize the economies of scale that are so important in the mass market or fund the research and development (R&D) to remain competitive. Even Ericsson and Nokia struggled against Huawei, with its Brobdingnagian $20 billion annual R&D budget, before the Western backlash against China. After its 5G revival, Ericsson drew a link between increased R&D investment and higher profitability. In 2016, it spent about $3.3 billion on R&D and reported a gross margin of 29.8%. Last year, when R&D spending had risen to $4.5 billion, its margin hit 43.5%.
Software will not change this paradigm as much as some open RAN proponents suggest. Even if people writing code do not require costly factories and assembly lines, software talent has become incredibly expensive. Nor do software giants have small R&D bills. Google poured $31.5 billion, about 12% of its revenues, into R&D last year. And Mavenir, a telecom software vendor and open RAN player, has previously outspent Ericsson on R&D as a percentage of revenues. In 2020, when it was considering an initial public offering, Mavenir revealed R&D investments of $89 million for 2019, representing 21% of sales, as well as an $81 million loss. Ericsson's R&D intensity last year was 18%.
Competition is already limited in some of these open RAN sub-markets, as Europe's telco incumbents pointed out in a report they issued last November. Their main worry is the lack of sizable European vendors across this landscape apart from Ericsson and Nokia. Even when non-Europeans are included, only four companies worldwide are classified as "major vendors" of general purpose processor technology AMD, Arm, Intel and NXP. Just five companies are identified as makers of distributed units (the boxes where baseband processing takes place) CommAgility, Dell, HPE, Kontron and Supermicro.
Operators are pleading for government funds to nurture open RAN startups, and yet their own venture capital (VC) arms have shown scant interest, as a Light Reading investigation revealed earlier this year. It has not gone unnoticed by the startups they call out in the report. "I have been surprised that while major operators speak about wanting to drive an open RAN ecosystem in Europe, their VC arms still remain wholly focused on software-as-a-service," said Adrian O'Connor, the CEO of Benetel, an Irish maker of radio units. "It has confused me to a degree."
Perhaps the clearest warning signs that a future open RAN market could lack much diversity are the targets flashed up by vendors. In the annual report it published last year, Japan's NEC said its aim was to capture 20% of global open RAN business by 2030. Rakuten, which uses NEC products in its own Japanese mobile network, wants a quarter of the entire RAN market, it told Light Reading at last month's Mobile World Congress (MWC) in Barcelona. If those targets are achieved, a massive chunk of the market will have fallen to just two Japanese firms. And other big players have similarly bold RAN ambitions. "Our goal is to be number one," said Woojune Kim, the head of global sales for Samsung's networks business, at MWC.
By the mid-2020s, open RAN will account for roughly 15% of overall RAN investments annually, according to Dell'Oro, a market-research company. And most analysts think its share will continue rising until 2030, when it may become the default. But that is not synonymous with diversity. A network that comes entirely from Nokia, using interfaces compliant with the latest specifications, could be qualified as an open RAN deployment.
Nevertheless, such widespread take-up of open RAN would evidently present risks to the incumbents. Huawei, Ericsson and Nokia together claimed 72% of all RAN revenues in 2019, according to Analysys Mason, a consulting and analyst firm. That means Asian vendors will not be able to realize their ambitions without hurting the incumbents. Ericsson has looked wary since open RAN came along, while Huawei remains publicly dismissive. Yet both are watching the market and investing, according to sources.
Nokia, without doubt, sounds the most enthusiastic of the three, although some operators and open RAN specialists are skeptical of its sincerity. "It could happen that in some networks we lose business to open RAN, but we think we can win more than we lose by winning from established big suppliers," said Tommi Uitto, the head of Nokia's mobile business, at MWC.
One European scenario, already playing out in the UK, is the substitution of open RAN for some part of the Huawei footprint. So far, Vodafone UK is the only brownfield European operator to have made a serious commitment to the technology, which will replace Huawei at 2,500 sites roughly 14% of its total (and more than 40% of the Huawei estate). Other Huawei sites are likely to follow. The main beneficiary, though, is not some group of open RAN startups but Samsung, already the world's fifth-largest RAN vendor, which is supplying both radios and baseband software in some areas. It might be open RAN by name, but it does not look very open RAN by nature.
Protectionism born of geopolitics will probably limit the choice for operators and the addressable market for suppliers. Dozens of Chinese companies are registered as contributors to the O-RAN Alliance, the group behind open RAN specifications. They include ZTE, the world's fourth-biggest vendor. Yet none can expect many invitations to pitch in the West. Globally, the RAN market may look diverse. In individual countries, it will probably not be.
Tellingly, the open RAN opportunists hungry for a big slice of the cake are starting to branch out and look more like RAN giants with the full range of products. This is not a surprise. NEC could not possibly hope to capture 20% of all open RAN sales with radios alone. Having concentrated on those initially, it has more recently added baseband software to its catalog. Samsung already had both. Even Mavenir, touted as a software specialist, has developed its own radio unit. "To make the radio commercially viable, you have to think end-to-end and the ability to integrate," Job Benson, Mavenir's head of product management, told The Mobile Network.
The irony will not be lost on traditional vendors. Open RAN is purportedly about breaking everything apart and specializing, not the in-house integration that happens at Ericsson and Nokia. If Mavenir starts thinking "end-to-end," what makes it so different? But the commercial logic is obvious. Suppliers with a range of goods can offer a better deal to operators, just as operators with fixed and mobile networks can lure their own customers to convenient, competitively priced product bundles. This possibly explains why Vodafone has turned to Samsung for both radios and baseband software.
Preferred ecosystems are also taking shape. While it has little desire to produce hardware, Rakuten has assembled a team of vendors, contributed its own software and begun a marketing assault through its new Symphony unit. Some revenues will obviously have to be shared. Rakuten has even said its app store of network software (Symworld) will be opened to rival products. What's important is that customers can interact with Rakuten Symphony much as they do with Ericsson or Nokia. 1&1 has already put the Japanese company in charge of a mobile rollout in Germany.
In any case, it would be a mistake to think the Nordic vendors do everything themselves. Under a 5G product refresh that started in 2020, Nokia flagged its reliance on the same US chip companies now angling for open RAN business. Broadcom, Intel and Marvell were all identified that year as important silicon suppliers. After it has added some of its own technical expertise to the mix, Nokia effectively plays the role of systems integrator, packaging up the various components and selling the final product to a telco. From a customer perspective, Rakuten is little different.
This reflects how operators tend to buy products through a few big firms rather than a jumble of small suppliers. Neil McRae, the chief architect of BT, doubts this will change much with open RAN. By reducing the number of suppliers, BT has been able to cut costs, he said at a press event last year. "I don't think we'd increase dramatically the number of suppliers we work with purely because it is hard to manage that integration and manage them all together."
"The big players love to be served by large multinational players," said Benetel's O'Connor. "I still think the operators will look to the big players for their solutions." For now, O'Connor's focus remains the smaller market for private wireless services, where new entrants addressing niche requirements have a much better opportunity. His thoughts on the mainstream telco market are revealing. "You could get back to having six significant players and I think that would be a very healthy position to be in," he told Light Reading. "Would it stay at six or would it start to contract again? Possibly."
Seven years ago, the European Commission (EC) invited telcos to comment on Nokia's proposed 15.6 billion ($17.2 billion, at today's exchange rate) takeover of Alcatel-Lucent. Did they object, fearing competition would suffer? "On the contrary, virtually unanimously, the customers who replied to the market investigation consider the overall impact on the RAN market to be neutral, with some customers even considering that the transaction may have a positive and pro-competitive effect," said the EC in a 42-page report it subsequently issued.
If nothing else, the feedback illustrates how swiftly priorities can change. Government intervention over Huawei has convinced today's operators they need more choice. Yet there is not much conviction it will either arrive or endure. One telco executive acknowledged there is concern the industry might go through a long and painful experience with open RAN only to end up back where it started, buying products from a few giant vendors. If that happens, and diversity is the main objective, this mission will clearly have failed.
- Rakuten's Amin targets 25% of RAN market in bid to unseat giants
- Nokia mobile boss lashes out at open RAN fakes
- BT takes aim at open RAN myths
- Open RAN might not save you much after all
- How telcos installed a RAN triumvirate they now want toppled
Iain Morris, International Editor, Light Reading