Between this week and last, a few announcements have emerged which highlight some basic tensions in how the telecoms industry evolves. They expose a few problems with the industry’s familiar idea of a mobile generation and raise questions about how we go about the “G” process.
So what happened?
Number One – the NGMN launched guidance on 6G Architecture and Migration Options, focusing on how operators can push for making a 5G-6G transition without having to build yet another network and hang another legacy albatross around its neck. The organisation is pushing for a limited set of migration options, with Multi-RAT Spectrum Sharing identified as the primary approach and other mechanisms considered where there is a clear requirement. The underlying principle is that telcos shouldn’t have to deploy equipment without a business case being established. All very much in line with NGMN’s approach from the past few years.
That was released not coincidentally the week before Number Two – 3GPP’s September plenary meeting, touching largely on Release 20. Release 20 is focused on 5G-Advanced alongside studies ahead of 6G exploring what the system should contain. 3GPP’s current 6G work includes potential areas such as AI, sensing, computing, distributed autonomous networks, migration and interworking, NTN, 6G IoT, security, QoS and policy frameworks, and immersive services – everything but the kitchen sink.
While that’s exciting in terms of opening up new use cases, every additional capability creates another argument for introducing new architecture, interfaces, management systems and equipment. In other words, exactly what the NGMN is hoping to avoid.
Meanwhile, Number Three – Verizon’s 6G Innovation Forum is putting together an ecosystem to work out what that network might actually be used for. This week the Forum announced new additions including AWS, Cisco, Intel, Keysight, MediaTek, NVIDIA, Palo Alto Networks, Rohde & Schwarz and VIAVI Solutions. They join Ericsson, Samsung, Nokia, Meta and Qualcomm. Verizon’s stated intention to approach 6G from use cases rather than start with the radio architecture.
The company has already tested some of those use cases over the summer. In a Samsung trial, a virtualised 5G network used changes in radio signals from six smartphones to estimate crowd density at a “major sporting event”. Samsung and Verizon described it as an AI-powered ISAC trial over vRAN. The use case offers “real-time intelligence for large-scale events, which could allow event organisers to monitor spectator flows in real-time, streamline gate operations, reduce concession line wait times, and proactively address overcrowding risks.”
Verizon’s AI Sports Companion prototype combined Meta glasses, edge computing, a small language model and NVIDIA hardware to provide real-time information to spectators – essentially people talking to and with their glasses.
None of this is a commercial 6G service, but it provides evidence about the kinds of workloads and interactions a future network may have to support… at least if companies like Meta can persuade people their glasses aren’t creepy.
Parallel Lines
So that’s three groups all acting more or less in parallel, and all with very different approaches.
3GPP operates on consensus. Frankly, with the number of stakeholders involved it’s a daily miracle that anything gets done. The people managing it are absolute heroes. However, it means that there’s a tendency towards standards bloat – not least because there’s still not a clear and uniform idea about what 6G would be for. Without knowing what is irrelevant or secondary, it’s unclear what should be thrown out.
Meanwhile the NGMN does have a very clear vision for the next generation. It’s a vision driven by the economics of telecoms players today, which essentially means minimising risk. Telcos were badly burnt in the 3G era, and the commercial promise of 5G is still in the process of materialising. Taking a gradual approach to deployment or implementation once the business value is clear is a rational approach. Again, it doesn’t actually take a stand on where the value of 6G lies – it’s more about steering the standards avalanche.
Verizon’s Forum describes itself as starting with the use cases, which is arguably closer to how normal businesses operate. They’re trying to identify where a possible 6G network would add value, and then figure out a business case or business cases from that.
Generation Collapse
What the NGMN and Verizon’s activities both point towards is a clear distinction between “deploying a 6G network” as we may have seen in previous generations and “adding 6G capabilities”, which – depending on business cases – may never be ubiquitous in the way 3G, 4G and 5G were envisioned.
What’s more, we’ve seen live implementations of ISAC “6G capabilities” being rolled out over 5G-Advanced, thanks to Samsung, Meta et al.
This is where the conventional generational model starts to become difficult to maintain. At what point has the operator started deploying 6G, if the capabilities associated with it can already be put to use?
On one level, hooray – perhaps the industry, writ large, can break out of a ten-year CapEx investment cycle and start to reimagine how it builds and offers services. That’s going to be exciting for some CFOs. However, it also creates its own challenges.
- Most network procurement processes are set up to be really good at managing cost and risk in relatively few very large purchases from relatively few vendors. If telecoms providers are going to follow the money in a more agile and flexible way then the procurement processes are going to need optimising for speed, small batches and flexibility, with changes of KPI and incentive.
- It also means that, instead of having a ‘killer app’, there’s going to be much more need for exploration and creativity in how, when and with who to deploy new capabilities. That means changing incentives, cost structures and back-office flexibility to enable rapid trials of solutions. Our reports have several examples of telcos missing out on working with important clients because they weren’t responsive enough or didn’t consider gaining a few tens of millions to be a worthwhile investment.
Verizon’s choice of partners also says something about the changing role of the mobile network: AWS, NVIDIA, Intel, Palo Alto Networks, Cisco, MediaTek, with Keysight and VIAVI to run testing and assurance. This looks less like a conventional telecom stack and more like an AI infrastructure stack in which connectivity is one of the essential components.
That could change the commercial proposition for operators. An autonomous drone does not need “6G”. It needs connectivity, positioning, sensing, compute, security and an application that can effectively combine them. Under the hood it also needs reliable handover and fallback, and may well need to handle SLA management, billing and settlement across different providers’ networks.
The network provider here is not necessarily providing the whole service, but some essential parts of it. Rather than ‘owning the customer’ it may be about owning unique elements in the service delivery chain and taking risk off the hands of both the customer and the business that does have that direct interface.
Smooth Operator
One weird element in all of this; while 3GPP is trying to figure out how to build 6G, we’re already seeing the emergence of the most exciting “6G stuff”.
ISAC is being trialled on 5G-A with Verizon, while Chinese telcos are managing robots in public over 5G-A. AI and machine learning elements are already being incorporated into 5G-Advanced; edge capabilities are, nominally, out there for the adoption; NTN integration is already being standardised and commercialised; and network APIs and exposure are being monetised today.
None of these capabilities needs to wait for a 6G launch date, though it might help with performance or economics.
That makes the likely migration path much smoother and more granular. An operator might introduce sensing in one part of its network while retaining existing radios elsewhere. It might add edge AI capabilities without changing the RAN. That would fit with the NGMN’s view that the network will evolve according to investment cases, but it makes for a really difficult trade-off.
The NGMN wants a manageable migration; 3GPP is investigating a much broader set of network capabilities; and Verizon and co are already trying to work out why some of those capabilities might matter, while actually implementing them without 6G.
So… as an industry, the ideal outcome means balancing two ideas at the same time.
- 6G needs to be different enough from 5G-Advanced to justify investment, but
- It cannot become so different that operators need a massive network rebuild to obtain the benefits.
Especially if we’re looking less at a network migration and more a migration of capability, this would represent a significant departure from the industry’s traditional model where the arrival of a new generation is the stimulus for massive new network infrastructure investment.
Not great news for network equipment providers. However, if the result is a 6G upgrade process where the technology arrives in pieces, that are advantages.
It gives operators a chance to test the commercial chops of new capabilities before committing to large-scale deployment. It gives enterprises more flexibility, and therefore offers the telcos more scope to work with them. And, for vendors, it offers a chance to put capabilities into the market at scale before any major “6G network” purchases have come in, just depending on how useful they are. So it could potentially smooth out the feast-and-famine ten year cycle.
If this kind of topic resonates with you, TelcoForge’s 6G Value Season of public and private discussions may be of interest. Get in touch if you’d like to share your perspectives.
Image by Canva AI with apologies to Gene Roddenberry
