Ericsson rewrites sales pitch in face of slowing traffic growth

Wearing a leather jacket that will last you the rest of your life sounds very Fight Club, but it's not in the interests of the average commercially minded clothes maker counting on holes and rips for return business. The same goes for mobile infrastructure. As networks squeak and strain under a torrent of gigabytes, their telco owners must regularly patch them up – to the satisfaction of product vendors – and more occasionally change the outfit. But for the last two years, those networks have looked worryingly resilient.

Analysts talk of excess capacity as telcos weather that gigabyte storm. Nearly two years ago, after a sales dip in North America, Ericsson CEO Börje Ekholm warned his customers they would eventually have to reinvest. "Operators can sweat assets for a couple of quarters but it cannot be done much more [than that] because of the traffic growth underneath," he said on a company earnings call in January last year. Sales have recently improved in North America, after customers used up older supplies, but they have fallen everywhere else. And there has been a distinct tonal change in the talk about traffic.

Seemingly for the first time ever in the mobility reports it publishes every six months, Ericsson has acknowledged a slowdown. "On the forecast side, we see continued mobile network traffic growth but at a slower rate," said Fredrik Jejdling, the head of Ericsson's networks business area, in the introduction to the report published this week. The sales patter is no longer about the volume of data and what telcos must do to cope. Instead, it highlights the value of new 5G network features.

All that is supported by a strategy built around what Ericsson calls the "programmable" network. Software that can be hosted on existing basestation equipment can pinpoint the location of users more accurately than GPS, boost performance as and when needed, extend battery life for smartphone users or power down mobile sites – to name a handful of the features Ericsson advertised when it recently launched 5G Advanced, the latest flavor of the current mobile generation.

Priority boarding

Ericsson deserves marks for dropping what was a somewhat negative pitch (spend or suffer) and trying a completely new approach. But programmability – while it sounds much more positive – may be an equally difficult sell (and so Ericsson is hedging its bets, which we'll come back to). Just as telcos have not been able to make additional money from traffic growth, so they may struggle to generate sales from the new software features that Ericsson is keen to license.

The one that looks most obviously monetizable is often referred to as "quality on demand." Thanks to a 5G technology called network slicing, a priority lane could effectively be reserved for a paying customer. This might be an Instagrammer needing a reliable service at an event, a business demanding a service level agreement or even a software developer. Using standardized application programming interfaces (APIs), said developer could drag and drop code that gives the application full access to a quality-on-demand feature.

The market for these APIs is forecast by Omdia, a Light Reading sister company, to be worth about $8.7 billion in 2029, up from just $161 million last year. For platform companies packaging up those APIs and making them more consumable, this sounds lucrative. Vonage, which Ericsson bought for about $6.2 billion in 2022, made just $1.5 billion in total revenues last year, for instance.

But $9 billion is small change for a global telco sector that makes about $1 trillion in sales each year. These operators will benefit only if programmability persuades subscribers – whether consumers or businesses – to spend more on network services. The problem is that adding speed, responsiveness and other capabilities to networks has not previously led to an increase in customer spending. In many cases, customers have spent less.

At Deutsche Telekom, an Ericsson client, monthly average revenue per user (ARPU) for postpaid customers in Germany has fallen from €23 ten years ago, long before the launch of 5G, to €19 today. Vodafone UK, another Ericsson client, has watched monthly ARPU for the same type of customer plummet from £31.30 in the final quarter of its 2013 fiscal year to just £17.20 at the end of the last one. Such trends do not instill confidence.

To the average consumer, "quality on demand" could smack of priority boarding on a budget flight, or the coffee shops that charge extra for the privilege of sipping while seated. The industry has failed to boost revenues by providing high-speed, low-latency connections that are "always on" for a mass market. Instead, it is restricting access to short bursts for those prepared to pay a premium. Customers without technical knowhow probably expect the 5G service they were pitched to be higher quality than 4G. Suddenly they learn there is an even higher-quality service, available "on demand" to the more discerning customer. Why is the standard 5G service no longer good enough for Instagramming Taylor Swift in concert?

Inevitably, there is also the risk of a backlash from the priority boarders. They are supposed to receive a better service, but when physics intrudes everyone arrives at the same time. A network slice cannot propel a weak signal in a high spectrum band through a solid thick barrier, or magically support connections when a basestation is broken.

AI versus Taylor Swift

If all else fails, Ericsson can always revisit its pitch about data traffic growth. The rate of increase, unfortunately, dipped to as little as 4% between the second and third quarters of 2024. That's down from 6% between the final quarter of 2023 and the first quarter of this year. Three years ago, the sequential increase between the second and third quarters was 8%. Ericsson, nevertheless, stands by a longer-term prediction that data traffic "is expected to nearly triple by 2030 compared to current levels."

To justify that, Ericsson devotes a chunk of its latest mobility report to speculation about the impact of generative artificial intelligence (GenAI) on network traffic. With AI "personalization" of services, users will be glued to their gadgets for even longer (oh, joy), it reckons. By this stage, humans will have lost the ability to even tie a shoelace without AI support. Fortunately, they'll be able to train their phone camera on the untied laces, upload that data to an AI assistant and receive help. These AI assistants, presumably speaking with the sultry voice of Scarlett Johansson, could be ubiquitous with new smart glasses, dispensing harsh dietary advice when they see you're about to scoff a doughnut rather than graze on greens. Uplink traffic could explode.

But Ericsson doesn't know that for sure. Plenty of experts think all AI processing will happen in large data centers or on gadgets and that mania for Instagramming Taylor Swift is likely to put more pressure on the access network than anything to do with AI. Even Ericsson is forced to admit that its predictions could be "offset" by developments in compression and device technology.

(Source: Ericsson, William Webb)

Ericsson's track record in forecasting is not great. Its June mobility report featured a massive downward revision of historical data after the latest numbers from telcos and regulators were fed into the model. William Webb, an independent consultant and former executive at UK regulator Ofcom, noticed Ericsson had more surreptitiously cut the data forecast that supports the latest report, too. The exabyte total for Western Europe in 2029 is 10% lower than it was in June. Ericsson has also cut the growth rate for 2024 from 30% in the previous edition to 20% in the latest, flattening an awkward bump in the growth curve as if squashing an unsightly zit.

Webb will enjoy and feel vindicated by those changes. For years, he has predicted an S-curve flattening of data traffic growth by 2027. He has yet to see anything in AI, extended reality or gaming to make him change his mind. In the meantime, those networks continue to look in much better shape than their vendors would honestly like.