Featured Story
Ericsson rewrites sales pitch in face of slowing traffic growth
Ericsson substitutes value for volumes in its patter after recognizing a slowdown in traffic growth – but it still bets AI will have a massive impact on the network.
Omdia and Dell'Oro weigh in with their latest assessments of the RAN market this year, and they're a gloomy read for kit vendors.
The bosses of Ericsson and Nokia got it badly wrong. Early last year, 5G spending cuts began to wound both Nordic kit vendors. Telcos in the US, where Ericsson and Nokia generate a fat wedge of their profits, had previously hoarded supplies and saw no reason to restock. But the "adjustment" would be "mostly completed" by June, said Ericsson CEO Börje Ekholm, anticipating "recovery by 2024." The message from Pekka Lundmark, his peer at Nokia, was similar. A year later, the market looks worse than ever.
Research from Omdia, a Light Reading sister company, reads like a horror novel for anyone in the radio access network (RAN) products business. Sales for the first three months of the year fell more than 20%, compared with the year-earlier quarter. Ekholm had evidently grown more bearish by the time Ericsson published its first-quarter report, describing one forecast for the full year of a 4% decline as "a bit optimistic." Now Omdia has revised down its own outlook after the worse-than-expected start to the year. Previously it had been looking at a 4% to 6% drop. Sales will fall 7% to 9%, according to its latest figures.
The other big analyst firm in this game is even more alarmist, calling the RAN market a "disaster" in a recent update. "The velocity of the deceleration is now simply unprecedented," said Stefan Pongratz, an analyst at Dell'Oro, in a LinkedIn post about his company's research. His preliminary estimates point to a year-over-year decline in RAN spending of up to 30% – and no less than 15% – for the first quarter. This range makes it the steepest drop that Dell'Oro has observed since 2000, the year of the dotcom crash.
Both analyst firms are expecting improvements later this year. But this is partly because the percentage changes will not look as bad after the sharp declines of 2023. "The market slowdown really intensified in 2Q23 which means that from 2Q24 the year-on-year comparisons should be less unfavorable," said Remy Pascal, an Omdia analyst, in emailed comments. Pongratz, in his LinkedIn update, said Dell'Oro is "still expecting some improvements in the second half."
The gut churner for Ericsson, Nokia and others is the possibility sales don't improve that much beyond 2024. North America's much-discussed inventory correction – a factor in recent poor results from companies like Cisco and Intel with relatively minor exposure to 5G – is only one reason for the drop-off in sales. Much scarier are the other potential factors. One, telcos aren't spending on 5G because it has not brought the sales growth they expected. Two, they aren't spending on 5G because their networks are coping with traffic growth better than vendors expected. Three, the rate of traffic growth is slowing.
Race to the bottom
The first is well documented and evident every time a 5G operator publishes its own results. Two cases in point are Vodafone and BT, UK-headquartered telcos that published full-year reports last week. Ten years ago, Vodafone was reporting monthly contract ARPU (average revenue per user) in Germany of €27.60 (US$29.92) for the January-to-March quarter. The figure this year is just €17.60 ($19.08). In the UK, it has fallen from £26.60 ($33.68) to £17.90 ($22.66) over the same period. BT, while generating much of its revenue outside mobile, has seen annual sales fall by £3.3 billion ($4.2 billion) since 2016, when it completed its takeover of mobile operator EE.
Yet there is no shortage of demand for the applications that 5G supports, as anyone who has recently sidestepped a smartphone zombie will appreciate. The problem is the race to the bottom in competitive mobile markets, where telcos have increasingly allowed customers to guzzle all the data they want for low and fixed monthly fees. This won't change in markets with at least three mobile networks and a healthy supply of piggybacking mobile virtual network operators. There will always be one company prepared to offer all-you-can-eat tariffs, pressuring others to follow.
The industry's great hope is that developers will cough up where customers aren't. The plan involves using application programming interfaces (APIs) – connectors between networks and applications – to give developers access to mysteriously unexposed and valuable features of 5G. A bank, for instance, could obtain the relevant standardized APIs to build an app targeting fraud (the idea being that the latest networks can more accurately pinpoint customer location).
Unfortunately, some of these features can be provided without 5G. Even if they cannot, numerous telco executives remain unconvinced developers will pay generously for APIs and auxiliary services.
By far the most risible of these APIs is one dubbed quality-of-service or quality-on-demand. At the click of a mouse, or download of an app, a customer's connection would be temporarily turbocharged, ensuring his Fortnite avatar avoids death by lag. But ignoring the fact most gaming happens on fixed lines, no API is a guarantee against basestation faults or signals failing to penetrate an especially thick wall. The more discerning customer may simply want to know why the standard 5G service doesn't already come with quality on demand. Better quality, after all, is what 5G was supposedly about.
After peak capex
None of this would be quite so distressing for Ericsson and Nokia if networks were about to keel over in a torrent of 5G traffic. And the somewhat uninspiring message from Ekholm and Lundmark is that telcos will eventually be forced to open their wallets or face network problems as capacity runs out. Yet there is still little sign of it happening. AT&T plans to cut its overall capital investment from $23.6 billion in 2023 to between $21 billion and $22 billion this year. Having spent $18.8 billion in 2023, Verizon is guiding for capex of between $17 billion and $17.5 billion in 2024. BT is over "peak capex," it announced last week.
Technology executives within telcos have also downplayed suggestions that services based on artificial intelligence (AI) will generate a new wave of network traffic. "It is not nearly as big a deal as people make it out to be," said Enrique Rodriguez, the chief technology officer of Liberty Global, at a recent London event.
"In general, the growth of network capacity has been doubling every three years," he elaborated. "From what we've seen over the last 15, there is always a different reason for that to happen, but that growth has been pretty stable, and we've not seen AI in itself be a significant disturbance to that growth when it comes to the network." There is an abundance of capacity on fiber networks and 5G also looks well equipped given the lack of any "killer application," he said.
Pongratz alluded to the issues of network capacity and traffic growth in his LinkedIn post. "In addition to the well-known coverage-related challenges with tougher comps in the advanced 5G markets, there are now serious concerns about the timing of the capacity upgrades given current network utilization levels and data traffic growth rates," he said.
Exponential, until it isn't
The idea that growth rates are slowing is politically inconvenient for many telcos. In Europe, telco bosses insist growing volumes of Internet traffic force them to spend heavily on network upgrades. Internet giants responsible for that growth should contribute a "fair share," they insist. At this year's Mobile World Congress, Christel Heydemann, Orange's CEO, used the word "exponential" to describe the growth of traffic on networks. But the data shows she is wrong.
One of the most compelling studies comes from Analysys Mason, a consulting and analyst firm. Published in August last year, it shows annual growth in cellular data traffic slowed from 90% in 2018 to 34% in 2021 and just 22% in 2022. For mobile handsets alone, there has been an even steeper decline from 104% in 2018 to 21% in 2022, says the report.
Omdia is guiding for another slowdown this year. According to its figures, cellular data traffic grew 37% in 2020 but is projected to rise just 20% in 2024. "The general trend of slower data traffic growth rate is not new, but it is true that the pace of that slowdown is more pronounced than it was expected say 18 months ago," said Pascal by email.
"This is therefore a factor in the slowdown," he added in reference to the latest RAN sales forecast. "There are others, for example we hear operators saying they plan to extend the lifetime of some equipment, sweating assets a bit longer and delaying modernization projects."
A problem for Ericsson, Nokia and other companies active in the RAN market is heavy reliance on Wi-Fi connectivity by smartphone customers in countries with expansive fixed-line networks. It is noted in a recent publication written by multiple analysts from numerous organizations, including Telefónica, Oxford University and Disruptive Analysis. Citing the Analysys Mason research, it takes aim at commentators who have "incorrectly claimed that Wi-Fi usage would be crowded out by cellular." In fact, Wi-Fi currently accounts for two thirds of all global Internet traffic, says the report.
William Webb, a former executive at UK regulator Ofcom, predicted a "near-perfect S-curve" in his 2016 book The 5G Myth, reckoning cellular data growth rates would fall to zero by 2027. Now the chief technology officer of Access Partnership, a consulting firm, he refers to recent Ofcom data as proof of the anticipated slowdown. Between the final quarter of 2022 and the same period last year, growth in the UK was just 15%.
"Disaster" seems an apt choice of word by Dell'Oro's Pongratz. Ericsson last year generated about two-thirds of its revenues in the RAN market, while Nokia owed 44% of sales to this sector. Each has laid off thousands of employees in the last year to protect thinning margins, and there are hardly any politically acceptable alternatives that could meet telco needs worldwide. For both the parsimonious and the squeezed, a lot is currently at stake.
You May Also Like