While there have been plenty of announcements getting media attention, there are some undercurrents which are less visible but which may make more long-term impact.
MWC this year has been intense for the TelcoForge team. From over 80 formal and informal meetings, recording 9 videos for clients, sitting in on press events and even the occasional conference session, there’s certainly been a lot to take in. It’s given us exposure to the big players, the smaller specialists, and even (gasp!) people who aren’t in telco! Here are some themes and issues which kept popping up across different contexts:
Most-used phrase of the week: “Technology’s the easy bit. It’s people who are complicated.”
That, or words to that effect, popped up throughout our conversations. There are some great innovations out there at all levels of the industry, but what will end up building markets or changing the industry depends on all the ‘messy bits’ that turn technologies into services.
Policy and regulation, pricing, the interests of customers, their partners and key stakeholders, the supply chains, market readiness, geopolitics, the availability of funding, getting attention and buy-in from the right people, the mindsets at individual, company and industry levels… all this tends to get disguised at MWC. The buzzwords tend to be things like AI, 5G, metaverse, NTN, RCS (a while ago) and so on simply because they are easy and straightforward.
While the headlines around MWC tended to be about the technology, most of the rest of this article looks at the ‘messy bits’ that were the undercurrents.
(Sidenote – I can’t help but wonder what an MWC would look like where the big buzzword was “brand value” or “customer engagement”).
So Long ‘Attention Economy’, Farewell Globalisation
Trust is a growing issue, and an opportunity for the seizing; trust that “digital things” are what they seem, that they will work as promised, and that they aren’t being subverted. It’s a currency in short supply these days. On a national scale there are questions of how to secure telecoms critical infrastructure; in AI, worries about how to avoid training models on their own exhaust and to avoid unwanted outcomes; for consumers, worry about deepfakes, scams and subversion, and with very good reason. There’s exciting work being done in authentication, identification, data verification/origin, security of all kinds, and it reflects an awareness that – while the internet players brought us a great deal of new experiences and opportunities – digital experiences hijack the kinds of heuristic we’ve evolved to create trust in the physical world.
The MEF’s global forum was a striking reflection of this impetus. All but one of the opening day’s sessions addressed the issue in one way or another. MEF has quite a special position bridging the telecoms and other industries such as media and fintech, so it reflected not only a preoccupation from telecoms players but also their actual or potential clients and partners in other industries.
This is a domain where ‘move fast and break things’ falls down but plays into the hands of the telecoms industry. I don’t think this is a minor issue, either. If the last 20 years were an opportunity to monetise attention, the next 20 are liable to be all about filtering out the unwanted, the bogus, the dangerous and the malicious. Assurance, verification, security, safety, resilience, regulation… it all feeds into this fundamental creation of a new digital environment which, for the end user, feels less like diving for pearls among pirhanas.
Closely related to this on a national and continental level, it was hard to miss the word ‘Sovereignty’ through the show. 18 months ago this might have been easy to dismiss, but geopolitics has changed fast and, in this area at least, it is having some very concrete impacts on the telecoms industry.
Towards the end of the last decade, India took a strong line on this with its initiatives on IndiaStack, Make in India, and stepping up on global standards development. Now it’s very clearly the turn of Europe to cry “A plague on both your houses!” to digital powerhouses whose friendship seems unreliable, and the outcomes are many.
Exhibitors throughout the value chain are leaning on their European-ness not because it scores brownie points with politicians but because customers want that and the policy environment rewards it. This week we saw it playing out in satellite, semiconductors, hardware, OSS/BSS and more.
There was some debate about whether this is a good thing.
On one level, it certainly stimulates competition. In many parts of the telecoms industry we see consolidation leaving just a few big global players that telcos have to rely on; in chip fabrication, in public cloud, in various network elements. It may not be a monopoly situation – not quite – but it limits competition and limits the incentives for radical innovation. This kind of regional shake-up could give smaller companies with different approaches enough oxygen to break through.
However, taken to extremes regionalisation could threaten the global nature of standards; and, while we might see new companies appearing, there is a risk that an extreme approach would mean that, rather than global oligopolies, we end up with regional oligopolies that still lack incentives to innovate but also lack the global customer base to do so. As one interviewee commented, “We say in Bangladesh that you can’t build a mosque with one and a half bricks.”
Double Agents
AI, and especially agentic AI, was all over the show this year. The conversation has definitely moved on from 2025, from wondering how we should use it to real implementations. In some cases, it’s about making services more accessible to end-users, masking complexity – for example, several test and measurement vendors have added natural language interfaces to their services so that anyone with a clear idea of what tests they want to run can create test scripts.
Meanwhile, Amdocs picked up from their personality-based avatar demonstrations last year to embed that technology in a whole layer, the “aOS” – a layer that runs across any OSS/BSS stack and enables engagement with end users in ways which mirror the kinds of communication they value. More about this specifically later, but two observations on the phenomenon:
Telecoms operators in many cases need not just transparent outcomes, but also deterministic ones. They operate in a business that rewards reliability – and, in some cases, punishes a lack of it quite severely. TelcoForge isn’t the only observer to question the tension between the deterministic impulses of the telcos and the probabilistic nature of advanced AIs. Credit to many vendors, they were able to explain how they are able to solve that problem by limiting the inputs and outputs available to the models. There has been real progress made here.
The thing is, this means that those vendors have added a layer of AI and then fudged it to imitate a deterministic set-up. Why not just go with a deterministic set-up in the first place, such as a more straightforward algorithm, and simplify the whole operation? It seems like using AI for the sake of it.
Secondly, while the use of AI to simplify the interface with complex technology is a good way to make the technology superficially more accessible, it makes the overall system that much more complex in reality. And that’s a problem, because…
Known Unknowns
The telecoms industry is a hugely complex system of systems. If that seems like an obvious point, it’s one which is frequently underestimated. To use one illustration from a conversation,
“I was in a GSMA working group recently looking at security, and we had one expert on this particular bit, one expert on this other piece, and so on. But we really struggled to get an overview of whether we covered everything between us, how the pieces came together and what the big picture was. And this was between experts, mind you.”
The sheer quantity of specialist knowledge needed to build and operate a telecoms system today is vast and growing. While groups such as the NGMN have called for 6G to simplify telecoms provision, which would be great, it seems unlikely when the demands being placed upon it for NTN integration, new capabilities and enhancements of old ones, new spectrum, AI-nativeness, reduced energy usage and more are all pulling in the other direction.
Putting a veneer of AI over the top can make this complexity more usable to people without the knowledge and skills to understand it in depth. However, people need to know this complex system in order to repair any breakages or to develop it further.
TelcoForge ran a survey in Q4 which, among other things, highlighted how aware executives are of the skills shortage within the industry. With an aging engineering workforce there is a risk that skills are simply going to retire out of the industry. Providing simple wrappers on complex technology is a good way to bring in young talent with a lower requirement for specialisation… but then, that lowers the demand for the specialisation to be applied until something goes horribly wrong, by which time it’s too late. We create a more immediately usable system while things are working, but one that ultimately becomes more complex and more fragile unless the industry and the companies within it spend more on in-depth training than they ‘need to’. That’s a hard ask of a CFO.
6G Questions
As alluded to above, 6G was present at MWC to a greater degree than in the past, but is still problematic. In part a clear business case hasn’t been made for it, while research organisations and vendors are cracking on with securing their IP and their position in whatever the next generation turns into.
There were some amazing demonstrations out there, especially at the NTT stand where for example a relatively simple robot engaged with the outside world through external sensors – one day, through Integrated Sensing and Communication. This demo takes the cost out of the robot, which is good, and puts it in the network, which is bad unless there’s a business model in place to make that valuable for the operator.
Overall, the basic challenge of 6G has been ignored – that 6G as a concept and in reality today is a wicked problem and isn’t being treated like one. There are ways to change that dynamic.
One would be to make the 6G development process much more of a transparent, iterative and cyclical process of feedback between those developing it and actual or potential customers, addressing questions both of technology and commercialisation. This is how to address wicked problems. That’s not how the current standards system works, though.
The other approach would be to hark back to the approach in 2G, which was not a wicked problem. Back then, telecoms providers were clear about what they wanted to accomplish, why and how. They gave a clear mandate to vendors who then produced the technology solutions to meet that. Not an easy problem to address, but not a wicked problem, insofar as the objective was clear and what ‘success’ would look like.
MWC would be the perfect time for a global operator body such as the GSMA to take a lead on this approach. However, according to telco insiders, “There’s a big argument going on there at the moment. The European operators don’t even want to talk about 6G, even while some of us are saying “Come on!”.”
While the NGMN has been doing so for several years now and has undeniably had some influence, it lacks the sheer clout of the GSMA. It’s been able to advise on desirable outcomes, but hasn’t been able to demand them.
The Road Less Travelled (for data packets)
While Orange, Telefonica, DT, e&, Vodafone, China Mobile, NTT, Viettel and others had magnificently visible stands, and were celebrated for example in the GSMA’s keynote programme, scratch the surface and alternative players were being targeted all over the place.
Solutions for MVNOs, private networks, satellite players, mission-critical networks, indoor providers and others abounded, not least by vendors who see new entrants such as these as a useful top-up. The national-scale telcos have well-understood buying cycles – and we’re not close to a peak right now. Meanwhile these alternative players are popping up like mushrooms, in some cases as sub-brands for the big players but often independently.
Mission-critical networks in particular had a subtle presence at the event, but much more of one than in the past; from a segment at Ericsson’s stand to bustle at specialist Kontron. The very particular requirements of these customers has previously tended to keep them on entirely different networks and segments of spectrum. However, that has also meant that they have missed out on the pace of development a global telecoms ecosystem could maintain. There is a convergence not only of technology capabilities – the easy bit – but also of methods to deliver them cost-efficiently with relatively little additional infrastructure spend. Lessons from pioneers such as Firstnet are working through into commercial propositions from vendors who have not only the ability to deliver the required end-state, but to work with their clients through an evolution from what they have today. That’s a big deal.
The Last Word
While the TelcoForge team had a busy time across the event, it was still a tiny fraction of what we could have seen. Was it statistically significant? Hopefully. But we’d love you to have the last word – what did we miss that also deserves discussion?
