Last week I was fortunate enough to run the 6G FORGE event at EURECOM, and the outcomes were really not what I had anticipated. This is a quick note on some of the observations ahead of the video footage coming out this week…
We had a very varied selection of around 30 speakers from different parts of the value chain – academics, R&D, standards of course, but also end-users from different industries and their suppliers, telcos, specialists in AI, blockchain, security, robotics, data, open networks and so on. The aim was to try and understand what value proposition 6G/Beyond-5G realistically offers to possible customers, and whether the process of 6G development can live up to that value proposition or if it’s on the wrong track.
Some things that jumped out at me as unexpected:
1) Spectrum was not a primary consideration.
There was certainly conversation about it, but either as a means to an end or as a moat to commercial competition. It’s hard to see the global spectrum landscape changing very much by 2030, and probably not by 2035 except in very high frequencies, so the impression I got was that the conversation here was largely pragmatic; we’ll have to work with what’s available, while making sure that developments in higher frequencies are applied judiciously to the right uses. The likelihood of rolling out nationwide coverage in mmWave or sub-THz is slim to none and Slim’s out of town, so we need to focus on the right tools in the right places.
One of the best comments was that we should stop trying to take spectrum from other industries that also need it, and instead start to work out how we can share better with them to deliver on what our collective users need.
2) Five+ years into the process, it’s still unclear who 6G is being standardised for.
Many stakeholders have a say in what they want another G to do – deliver better coverage, more sustainability, more data faster, cheaper and to more diverse devices, plus AI, plus sensing, plus integrating NTN, WiFi, etc… and to be more secure and less complex than ever before. So there’s a conundrum for a start.
It’s probably possible to do much of this, but probably not all in one place and time. Instead it’ll be a host of different elements being applied in different places. Bert on the street doesn’t need assured latency for his podcast listening, but industrial robots might in order to offload computation effectively. That means moving away from thinking of a network as a monolith.
And that’s fine, as we’re seeing the erosion of the 3G-era national players. Increasingly, specialists in private networks, neutral hosting, networks for railways, public safety, indoor coverage, smart agriculture and so on are popping up like mushrooms. Each of them has different requirements from their customers, so they’d be prioritising different elements.
At the moment, though, there isn’t a lot of attention given to these non-national-scale players in the development process, despite the fact that – if the intention really is to move away from mass B2C offerings – in all likelihood these will be the first movers in actually deploying the elements their customers want or need.
3) The standards process is changing (good) but maybe not in the way customers need.
Last month’s 3GPP meeting had a wide variety of input from non-telco participants, which is definitely good… but their impact on decision-making within the organisation is very limited, which is bizarre if these are representing the customers.
We heard several times from people – mainly end-users and the people that sell to them – that enterprises want solutions when they have a problem. They can’t wait five years for standards and testing, then more time for things to be turned into products. Much better to have something now. What’s more, they’re okay if there are difficulties or limitations with the solutions. As long as they know what they are, they can work with that. Better to have a bad solution and iterate on it in real life than spend years in theoretical discussions.
What does that mean for the process of standards? Should we start thinking about individual work items as mini-solutions, get them out the door and into use as and when possible, as opposed to waiting 18 months for an official 3GPP Release? Should we put proposed deliverables onto Github for people to experiment with… and thereby start to build a wider community of developers?
I don’t think we have clarity on all the answers, but there seems to be an awareness within the standards organisations that there’s a need to do things differently and they’re taking steps to make that happen.
The big challenge will come from the relatively few large stakeholders who like 3GPP the way it is because it works for their IP. However, they face a fundamental risk here: either they adapt, or else customers will find other ways to get done what they need to do and render those large stakeholders irrelevant in solution delivery.
4) Monetisation opportunities don’t lie at the radio or the edge.
While there are certainly use cases for joint sensing and communication, the radio is unlikely to become a major revenue generator. By contrast, different conversations focussed on the potential of continuing the current direction of software-based networking with open APIs into different elements of the network, whether along the lines of Camara/GSMA’s Open Gateway, the kind of xApps and dApps available in Open RAN, or other elements. This gives the telecoms providers elements that can be monetised in other developers’ solutions and services, rather than trying to own the whole service.
Note – this isn’t particularly “a 6G thing” as it picks up on current trends, but it shows where both customers and telcos can see positive steps forward.
At the same time, several speakers discussed the difficulty of setting up 5G in industrial and enterprise settings, which has been putting off large-scale adoption. Using automated systems and AI interfaces to simplify how and where to place antennas, to provision and configure systems and more would offer great value in removing the pain of implementing telecoms. Wi-Fi might be less secure or more buggy than 5G but it’s easy to install and fix for IT staff.
Once more, this is a business issue which is not new for 6G; indeed, it’s been the same for generations. However, we’re finally in a position where technology may be able to mask that complexity for end users.
5) AI for telco is not a money-maker.
Several talks addressed AI, but there was a clear duality. AI usage within the network is currently focussed on optimisation, whether for channel management, energy usage or other elements. That’s likely to be the case for the foreseeable future. Some very interesting discussions focussed on the use of LLMs to convey intent rather than data, which could considerably reduce the overall quantity of data being transmitted back and forth.
Meanwhile, where can telcos make money from AI? By turning the network into an asset for other companies’ AI to use, according to some. However, this did not go unchallenged. If AI is the new jewel in the crown for many companies, would they be happy having that sitting on somebody else’s hardware? How would insurance work in these cases, or SLA management?
6) While we’re in a conservative industry, there’s more radical thinking going on than one might expect.
I think people realise we’re at an inflection point where the current state of affairs isn’t working terribly well for many of the major stakeholders. While national-scale players have to take time to recoup their billions in investments, newer players are finding niches where they can use the latest software and hardware to do things very differently.
What, for example, can we do by taking authentication off a physical SIM, by having a network core running on a Raspberry Pi, and experimenting with the use of unlicensed spectrum? Could we see that being cheap enough to enable tiny-scale networks or even personal networks that use AI to negotiate with and connect to others around them? While this is anathema to major players, the capabilities are out there, certainly in some countries. While this example might not take off, it’s clear that there is scope for some radical innovation in business models based on the latest generation of technology, and more so as new capabilities are developed.
In part this goes back to the question of who we’re building 6G for and what it might look like. While the centralised organisations are aiming to create something universal in scope and, they fervently hope, geographically unfragmented, the applications of the technology are liable to feel very different depending on where you are, what you are and what you’re doing.
7) If you think manpower shortages in telecoms are temporary, think again.
Even while hosted at a seat of learning for the telecoms industry, several people voiced concerns that, as a global problem, fewer and fewer bright young technologists are interested in the telecoms industry. Not many people see the prospect of taking an idea through different processes for a decade or so as appealing. After all, they could be working in IT, AI or the internet where your ideas might be realised in months.
As a basic necessity, the telecoms industry is going to have to reshape how it operates and how it appeals to new staff. In the 1980s and 90s the telecoms industry was the bright young thing making waves and causing disruption, and that drew in talent which is now very senior. Reshaping internal processes within companies could help with some companies’ appeal, but we also need to look at the industrial infrastructure of standards, industry bodies, consensus and more which has taken us this far. If it’s struggling now, it might not be fit for purpose in a few more years.
All in all, the conversations pointed to an industry that’s ripe for reinvention and disruption. Topics which were front-and-centre even a few years ago were secondary here, while traditional boundaries between different technology and business areas are clearly eroding.
We asked the audience “Is the 6G process on the right track?”. At an occasion where you might think all the enthusiasts were present, the answers were almost evenly split between yes and no. If nothing else, that element of doubt has to be a healthy sign for the industry as a whole.
