DBTT Episode 8: Wi-Fi Now, FTTH Council, FutureNet & thoughts on Broadband, Sovereignty

In this episode (recorded on April 27), Ben & Dean discuss Wi-Fi and the recent Wi-Fi Now Global Congress in California, takeouts from the FTTH Council conference in London and what's next for the fibre & optical sectors, thoughts on Network Automation and the FutureNet summit in London and quick takeouts from last week's Unthinkable Lab event on Telecoms Sovereignty.

00:30 - Wi-Fi trends, including Wi-Fi 8, AI + Wi-Fi

06:00 - US policy & regulatory about Wi-Fi, including ban on international router imports, thoughts on 6GHz spectrum

08:57 - Benoit's thoughts on the FTTH Council Europe event in London & the wider industry. What happens to the industry when fibre rollout is almost complete? What about Africa?

13:25 - Optical LAN, FTTX in-building, copper wiring/Ethernet replacement and Photonics

19:40 - Fibre uptake trends in Europe

24:00 - FutureNet and Unthinkable Lab outcomes. Trends in network automation, OSS / BSS, AI+telecoms trends, sovereignty and 5G standalone.

29:40 - What's next? Ben's partnership with Point Topic on FTTH takeup acceleration, Dean's May event schedule.

31:40 – Book Ben & Dean for your event! Get us on stage at your conference, or facilitating at your private meeting or executive offsite. Connect & message us on Linked.

Links:

Wi-Fi Now Global Congress: https://wifinowglobal.com/usa-2026/

FTTH Council Conference 2026, April 14-16: https://ftthconference.eu/

Dean's FutureNet commentary:

https://www.linkedin.com/posts/deanbubley_telcos-edge-sovereignty-activity-7452275477878132737-7aVy

https://www.linkedin.com/posts/deanbubley_5gsa-5g-neutralhost-activity-7452614197273849856-AODk

Unthinkable Lab on Telecoms Sovereignty, April 23: https://www.telecomtv.com/content/unthinkable-lab-sovereignty-in-the-unthinkable-age/

In May, Dean will be at Connect(X), Canadian Telecoms Summit, TelecomTV DSP Leaders  & Netaxis Inspiration Day. Click here for details and to arrange a meeting.

Transcript

SPEAKER_00 0:04

Hello and welcome to Dean and Ben Talk Tech episode eight, is it? Eight? Yeah, number eight. We're doing well almost into double D. Yeah, good. Hi Dean, how are you? I'm good, thanks.

SPEAKER_01 0:16

Yeah. Pretty busy. Yourself?

SPEAKER_00 0:19

Yeah, well, quite busy as well. I I had a bit of a break, so um nice and refreshed now, which is good. So you were in the US a lot uh lately?

SPEAKER_01 0:30

Yeah, I've had a couple of trips, and in fact I'm going back again next week. Um but I'm particularly a couple of weeks ago I was at the uh Wi-Fi Now uh Congress in uh Mountain View in California, at the which is always a fantastic venue called the Computer History Museum, which is which is interesting as a visit in its own right, but it's a good place for conferences. And that was over three days. The first day was looking at sort of a long-range version of Wi-Fi called Halo and another wireless temperature called UWB. Both interesting. Halo is important for some IoT use cases and was more advanced than I'd expected. But the two main days um were focusing on Wi-Fi, both enterprise and residential settings. Probably the the big story which was being pitched by many of the speakers was looking forward to the next generation of Wi-Fi 8. Now, it's Wi-Fi cycles arrive much faster than mobile cycles, so we're only just starting to deploy Wi-Fi 7, and the standards for Wi-Fi 8 are being formulated by the smurly sort of pre-standard chipsets, and some people thinking that actually it's all iterating too quickly. But the interesting thing for me was the argument that Wi-Fi 8 is not about maximum speed, but it's essentially being well, they didn't use the phrase AI native, as that's very associated with um the mobile industry, but what they're pitching um Wi-Fi 8 is is Wi-Fi for AI, both Wi-Fi for AI and AI for Wi-Fi. And it's a bit of a similar story to what I've seen elsewhere, but it's actually a bit or in some ways more compelling. Um I think that the um the the Wi-Fi, the AI for Wi-Fi in terms of improvement of design and planning, uh performance, things like beamforming, there's a lot of sort of clever stuff that's going into the standard to optimise for low, low, not just low, but deterministic latency um and reliability. And some of that's AI driven, some of it's more sort of guts of the radio. But it's it's interesting that they're very much pivoting the technology to be more predictable and reliable um even in congested settings, which I thought was important. Um the other side to this is is putting Wi-Fi use for AI workloads. Now, some of this I didn't believe. There's an awful lot of talk about oh AI is going to need ultra-low latency. And I'm like, well, really? I mean, I'm uh if I run a query on on Chat GPT, it can take 30 minutes. Yeah, and I've put and coding a lot of the value in AI is for things like coding, which might run again for hours or overnight. So there's bits of AI.

SPEAKER_00 3:20

The half second of expressing that to the network is not where you're gonna challenge.

SPEAKER_01 3:25

Exactly. So I think that there's a there's a a sort of and I see this across telecoms, there's the sort of automatic, you know, like AI is growing, therefore access network, therefore low latency is essential. And I think there's a few missing pieces of that argument for certain things, mm, you know, in terms of the internal processes, if you're doing reasoning, if you're doing training, it's in microseconds to get data from the memory to the GPU training cluster, whatever. Sure. Um, but on the access network, I I don't think it's necessarily for for for chatbots or you know, whatever. A lot of this is like projecting that we're going to use AI for augmented reality and real-time analytics of video, which is is it's true on a factory production line, but for sort of someone doing TikTok videos or something like that, it's it's milliseconds don't matter.

SPEAKER_00 4:17

Also, even in private networks, um it there's a cost to doing ultra-low latency in terms of spectrum allocation and all that, which means that you don't want to have that as a default setup because then you're maximizing your cost all the time. Um, so you're only going to really dedicate those resources to applications that absolutely need it. Absolutely. But having said that, I can see that there's a certainly in the in the B2B space, there is a um you know, push and pull there where Wi-Fi does not want to be displaced by mobile private networks everywhere, and so they do have to ensure that they can handle these, you know, very specific applications.

SPEAKER_01 5:01

Yeah, certainly on the enterprise side, yeah, that makes a lot of sense. The the interesting one for me though, for the residential Wi-Fi, is the smarts you can put into a home gateway or router, um, which allows you to do stuff not so much low latency but higher privacy, and if you can localize stuff like you know, monitoring your um security cameras or potentially using doing motion detection, running other applications on the home gateway, potentially with a fair amount of privacy, we can debate whether the service provider is involved or not, then as well as latency, it it sort of is more on-prem, and also from the service provider's point of view, it's your electricity bill, not theirs.

SPEAKER_00 5:48

Yeah, of course. Um so so the other thing I wanted to ask you about is because last time we spoke about some of the um seemingly a little strange US policy announcements, which I imagine, since they concern home routers, would uh would would be discussed at such an event.

SPEAKER_01 6:09

Um it was quieter about it than you might have expected, probably because there's a lot of discussions behind the scenes. The FCC didn't send someone to the event, which they have done at various previous ones. So there's clearly, you know, sort of um let's say a vibrant debate. What I did pick out was a couple of the component suppliers um and sort of you know, people doing radio chips and things like this, we're very much like had uh Stars and Stripes made in the USA banner. So there's there's already a sort of interest in you know what does it take to make all this stuff domestically. There's been, as far as I know, three waivers issued, um two of which came out during the show. So um uh Antran, Netgear, and now also Amazon Eero have been given waivers at least until next year. Um so there's clearly a process behind this, and I saw the FCC now has an FAQ up on their website of like what does all this mean? So, for example, it doesn't cover your hotspot function in your phone, it doesn't cover enterprise Wi-Fi, which has generally been a bit more cybersec uh checked and verified, you know, it doesn't cover critical software updates. There's a f they've actually obviously had quite a lot of pushback on this, a lot going on behind the scenes. What what there's also was a little bit of an undercurrent about though, that the FCC is also still looking on the spectrum side, about you've got the CTIA starting to come out again, along with GSMA saying, oh, maybe you need to have another look at 6 GHz unlicensed spectrum. Um I've got some stuff I've written where actually 6GH is being used by Wi-Fi already, including the upper part of the band. And there was also surprisingly um something which actually got kicked out, which is on the sub 1 gigahertz, the 900 megahertz section used by this technology, Halo, and also things like LoRa and other low-power IoT networks, there was a desire as a by company called NextNav to use that for um high power 5G, which would have killed a big chunk of the IoT industry, including stuff used by utilities and critical infrastructure. Um, so that's been kicked out by I can't sure it was the FCC or the super or one of the courts that kicked it out, but that that that not sub one gigahertz is really important for lots of uses, including broadcast and microphones and bits of 5G and future 6G. Um, but it's good that at least that wasn't something that suddenly got thrown into the um you know we need to find spectrum to auction um sort of uh um yeah, grinding machine in the US.

unknown 8:53

Okay.

SPEAKER_00 8:54

Good. Well why while you were um in the US, I was in London for the Fibre to Home Council event. Um and um I I have to say I'm not entirely sure why yet, but I have to say it felt like it was a lot quieter than uh some of the previous uh iterations. It could be in part due to the venue because Excel is such a vast place that you know even um significant events may seem a bit smaller in proportion to you know the size of the venue. But I do think there was uh a bit of a question mark um around where the industry is going. Because if you think about it in Europe, uh there's a number of markets now where access upgrade is you know virtually done. Um we're talking about the last maybe you know 5% of homes, um, which in volume is going to be insignificant in light of what was done before. Um, and there are countries where the growth is just not happening in large part because uh the key market players are not keen to invest all that much, and so I think uh part of the industry is really questioning how they keep you know their revenues at the level they currently are, let alone increase them. Um, and I think there's a number of uh paths there, but none of them super simple. So obviously, the most obvious path is well, XGS Pond, which is the 10 gig symmetrical pawn, is now more or less industry standard, uh, which doesn't mean it's been deployed everywhere, but it means if anyone is deploying now, they're deploying um XGS Pond natively, and any upgrade cycles happening would go from you know the legacy G PON to XGS Pond. Um but the next wave is either 25G, which is kind of what Nokia pushes for, or 50G, which is what kind of the Chinese vendors and a few others push for. But um I have some doubts as to you know the significance of that change in the next few years, um, simply because it seems to me that in light of overall internet traffic slowing down its growth to a point where we can now more or less anticipate it being flat by the end of the decade, um, is there really going to be demand for super high speeds going forward? So this is kind of um you know a parallel with the point you were making about Wi-Fi now focusing on reliability and and other uh metric and aspects of what a good quality of experience may be. Um and uh so so so that's one path, but I think it's not going to drive the levels of revenue that we have seen in the industry previously. Um another path would be um, at least for the vendors, would be expansion outside of Europe. So leveraging the know-how in markets, you know, that are still where things are still to happen. Um but if you look at that kind of coldly, it's gonna be primarily developing markets, and for European players is probably gonna be primarily Africa, and a few of these players are already positioned in Africa, and I'm not sure that the market dynamics there are at the level yet where it could sustain a substitution of revenue from Europe to Africa.

SPEAKER_01 12:51

Yeah, presumably it's gonna have to be some countries in Africa, or yeah, like I could imagine sort of Morocco, maybe parts of you know, Egypt and yeah, a few others, you know.

SPEAKER_00 13:01

But uh, yeah, well, obviously South Africa, Kenya, maybe Nigeria to some extent, but yeah, it it is going to be much more patchy, and certainly the prospect of you know universal coverage or something approaching universal coverage is is definitely not on the cards. Um and then there's a third path, which I think is a really interesting one, but it's one that the industry has been struggling with for a while now, which is optical land. So ironically, fiber in the axis is now a given, even if it hasn't been deployed everywhere, but um inside businesses we're still relying on copper everywhere. And um optical solutions are actually very effective and cost effective, much more so than your legacy Ethernet stuff, and the volume of business that could be derived from that is actually enormous, it's really, really huge. The problem is this is a market that Cisco has preempted, and you know, for the last 35 years or even more, this is all Cisco, and so no one has managed to really go head to head with Cisco and build the kind of distribution network that Cisco has, which is one of their great assets. Um and so, you know, basically, if a company gets to a point where they say, Okay, our Cat 5 is now insufficient for our needs, well, they're just going to deploy Cat 6 instead of looking at alternative solutions that might be, you know, longer lasting, less energy uh demanding, all of the stuff that optical fiber could deliver. So I think there's a big question mark there, which is I'm sure there is a big opportunity, but I'm not sure the industry has sussed out ways of actually leveraging that.

SPEAKER_01 15:01

Yeah, no, I mean it's it's it's a tough one. I mean, you've also got incumbents, you know, whether it's Cisco or HPE or uh Extreme, and there's a there's a list of other other ones. And it's it's the issue is that that ultimately it's still Ethernet, and there's a big link to the channel, to skills, you know, there's millions of people who know how to put Cat 5 or Cat 6 or whatever uh cabling in a tray and around around a building. Um the number of fiber installers is smaller, you've got more dependencies on the connectors, it's you know I mean obviously you know you've got areas where it's essential, and so I mean I mean to be honest, the the f if you're talking about copper replacement, where it is starting to happen is in the data center um of co-packaged optics between um you know rather than pluggables, um, and that's I mean there's you know massive shifts in terms of investors at the moment, but the the use case there is you know actually terabit level networking between GPU clusters and high bandwidth memory and all the rest of it. Now, maybe over time, you know, what we will see the same way that the sort of enterprise and data center LAN eventually filtered down to smaller businesses, maybe you'll see the optical and photonic side of it go from hyperscale data centers downstream. I'm not sure whether the grassroots, you know, you want to sort of connect your PC to the printer um stuff uh at an enterprise level is gonna start with you know optical connectivity and stringing fibre around the the office for any time soon. I think it's gonna filter down rather than filter up.

SPEAKER_00 16:49

Well, I think there's there's kind of some use cases are obviously more compelling than others. So you know, the hot the hospitality industry, hospitals, um, you know, there there's there's I think there are fragments of the market where the business case is easier to make. The question becomes how centralized can those decisions be? So, you know, I have no idea how many hotels there are in Europe, but but it's a huge market potentially, if you were to, you know, sign up a deal to refurbish even a tenth of those. Um, but the question is, you know, actually practically, how do you do that? How do you make that happen? The other angle is the real estate angle, I think, where essentially uh if you look at this from the point of view of real estate managers as opposed to the companies that might be actually you know leasing or renting the space, um, there are other considerations. So, for example, um when uh when each new generation of Ethernet is deployed, the cable of the previous generation stays in place. It's not used, but it stays in place. So there's actually a very significant weight of copper cabling in the false ceilings uh of most buildings, and in some places, yeah, in some places this has started to become an issue because it's a weight bearing on the entire structure that's actually not negligible. So, you know, there may be other incentives for real estate to say, okay, we need to be modernizing from that point of view. The other thing, of course, is there is some space to be gained from passive um optical deployment because you don't need active routers on each floor, you don't need to there is a problem, there is a problem though, which is one of the nice things about copper Ethernet is you can do power over Ethernet.

SPEAKER_01 18:45

So you can actually power the endpoints like Wi-Fi routers or it's in small cells. Whereas essentially you may if you can put optics in it, it may just mean you need to put more electrical outlets and electrical wiring instead.

SPEAKER_00 18:58

So, I mean, uh I don't want to overstate the case because all I've seen is business cases pushed by vendors who obviously have a vested interest in showing these business cases are positive. However, uh at the event I did have the opportunity to speak to a few companies, um, and you know, they seem to think that it was mostly beneficial. So there there are some trade-offs for sure. Uh, I do think that the overall case is good. The problem is more one of distribution and go-to-market, basically. How do you how do you make that happen en masse? And I think it's a it's a very interesting challenge to be had, but it is a challenge.

SPEAKER_01 19:39

Um so one question I've got for you. In terms of, I mean, I I would have liked to have gone to the FTTH event myself, especially as it's uh it was in London, but I was I was away that week. What was your sense in terms of uptake and trends in converting you know properties past whether it's sort of residential or MDUs or enterprise to actually using you know fibre? Because I know some markets are it's a done deal, but others like the UK, it's sort of fairly grindingly slow to actually get people to to switch.

SPEAKER_00 20:11

Yeah. So I mean, I think we have discussed this in the past, and essentially in markets where uh fiber to cabinet was rolled out quite effectively, they're all struggling with uh fiber to home take up simply because you know, whereas in France and Spain, your your arbitration as a client was do I want to stay on my 10 megabit DSL or do I want a gigabit fiber for you know the same price or a slightly marginally higher price? That's a relatively easy decision to make. Um, whereas if you have you know a hundred or even two hundred, as is sometimes the case in Germany because they've got bonded pairs of copper FTTC, then your gigabit at the same price or higher, it doesn't look very appealing. Um, so it's much harder to transition customers over to fiber in those markets. Now the trends have been accelerating slightly, but it's not massive. And this is where one of the one of the topics that keep keeps coming back that does irritate me is you know looking at copper switch-off as the silver bullet to that equation, which I think is completely absurd because the timing is entirely wrong. If your investor is unhappy with your lackluster take up now, they're not gonna wait 10 years for you know copper switch off to happen. Um I I I'm gonna come back to this if you're okay at the end of the session, uh, because I I'm talking to various potential clients about a solution that we uh want to offer. But what I would say is the following I think that the topic of take up is really. Down to commercial strategies by the various players who have deployed Fiverr. And I think by and large, there's been in the last 10 years an overarching focus on deployment as opposed to selling. And so I think not a hundred percent, but a lot of the you know potential solutions there are going to be about operators doing the right thing from a commercial standpoint and actually, you know, taking the topic seriously as opposed to kicking that can down the road or uh just putting their hands up in the air and saying, well, people are not interested. I think it's it's a commercial challenge more than anything.

SPEAKER_01 22:40

Yep. No, I I can I I I mean I certainly see that one of the people I've spoken to, you know, there's obviously been a sort of prob I think they probably pushed down on the the build pedal for a bit too long and didn't switch into product uh yeah, particularly not not just sort of sales marketing, but actually thinking about products and bundling and all of that stuff. Um you know, and actual service creation and innovation, particularly in markets where there's a lot of competition.

SPEAKER_00 23:10

Well, uh what I would add to that, and I agree with that, but what I would add to that is also that I don't think it's something that you I mean, obviously you have to have some built in order to have something to sell. But the best time to sell is actually when you've just built. Because people know you're there because you've been, you know, digging their streets and and your vans have been visible and all of this. And doing that and then waiting three years to start commercialisation on the dubious um you know justification that uh you're not at scale, I think is a big mistake. You actually need to be selling the minute that you know you're ready for service, and uh obviously, you know, when you're ten years late on that, that's not going to be a solution. Yeah, but I think it's a big part of the problem.

SPEAKER_01 23:58

Yeah. Okay, um, thanks for that. Um a couple of things. So last week was pretty busy for me. There was um uh FutureNet uh in London for a couple of days, which I attended, and I moderated a panel on APIs, and then I have my own uh unthinkable lab on sovereignty on what it means for telecoms. The two sort of were there was a fairly strong overlap. So I'd say the FutureNet, FutureNet's a fantastic event. They're they've they've run in London for a few years, and there's a couple of others in New York and one in Asia or Middle East now as well. And it's it's it's essentially the main focus is on um automation for telcos, particularly sort of OSS and to a degree BSS, so the back office systems, and also it's sort of historically the NFV and managing the network virtualization parts. And over time it's absorbed a bit more of stuff on the radio network and the fixed access network and AI and you know APIs. Um, I would say the one this last week was probably went a little bit back to its roots on automation. There was a lot of discussion about where are we with you know, quite autonomous networks, and the the TM forum has got a scale that matches autonomous vehicles, so there's a lot of people saying, Are we at level four? with which means you essentially have the human sort of administering the loop. There's various on the loop, in the loop. I'm not gonna go through the details of it. Um and the the general story is it's progressing. There's a bit of wariness of doing full lights out automation because of concerns over um you know well essentially sort of risk and proof points before you completely let the robots run the network. But certainly on greenfield networks, there's there's more interest there on certainly on the deployment side of it. You've got companies like Racket and Symphony, which are which I've been advocating that for a while. There was probably much less talk of the um we'll let AI run the network and then use any spare horsepower to run customer workloads. That was much less of an emphasis than I might have thought. Nvidia wasn't a sponsor. There were booths from uh I think AMD, chip manufacturers there, also Broadcom, but that's probably Broadcom wearing its VMware virtualization hat rather than the silicon side. Um what I thought was interesting was the probably the first panel of uh CIOs and other C levels talking about AI and uh networks, and they basically said we're not seeing any growth in uplink traffic worth noting yet, maybe a small amount of growth from a small base. Um we're interested in edge compute, but more for reasons of sovereignty and offering sovereign services rather than the usual sort of AI latency robotics, physical AI stuff. They're watching it, but it's not there happening yet. And the the the importance is for telcos doing AI internally, if they want to have an external AI services play, there's a bit of a move towards setting up separate what are called AI factories, almost as like a sort of a develop development house uh in parallel, rather than sort of thinking about reusing the same infrastructure. And that to me makes a lot of sense. I mean, it's it's hard enough for operators to deploy AI to improve their network, build an operation, without thinking about am I putting the right resources in place with developer tools and support and all the rest of it for customer workloads. And I think separating those two out I think makes a lot of sense to me. Um so I thought that the FutureNet was very interesting. I ran a panel on network APIs, which which was much more um in the background this year. It's sort of it's been going on for two, three years. We've got anti-fraud APIs, we're still quite you know slow on getting the hard network APIs of quality on demand, they're popping up in little places, but it wasn't an emphasis at all for the overall show beyond my panel, I think. Um and then there was also one very interesting panel which I wrote a LinkedIn post about about 5G standalone and the fact that some operators are still essay skeptic, um, basically saying, you know, if we were on a 10 euro ARPU, we're not sure we can squeeze any more new services, even with the fancy feet features from standalone. So is it worth it? Um and and I thought that was a bit of a wake-up call, whereas you've got some operators that are very much, oh, 5G standalone is real 5G, it's fundamental to everything we're doing, it's a foundation for 6G in the future. Whereas others are like, well, show me the money. Um and then lastly, the and I I I ran last Thursday the um fifth in the series of my Unthinkable Lab uh workshops, which are um you know small groups, uh 20-30 people. I run them with another um uh colleague of mine, Andrew Collinson, supported by telecom TV. Um this was about what does um sovereignty mean for telecoms. Um I wro again I wrote an update post, it was a very interesting day. Some of the details are under Chatham House rule. We'll write up a summary report. So I think next time we do a podcast, I'll refer to the the report when it's published and and uh give a link. I'm not gonna rehash it all right now, but it was that there was a lot of insight, and yeah, sovereign sovereign sovereignty for telcos is different to the big picture digital sovereignty that politicians because when politicians talk about tech sovereignty, 90% of it is about cloud, AI, and chips. And networks are sort of assumed, and it's it's a bit more subtle. I agree. I agree. Right, good. Um okay. What's what's what's next?

SPEAKER_00 29:46

Well, I I said that I would talk about uh take up um quickly uh at the end. Basically, um I've been working with a company called Point Topic that specializes in um geographic geographical data. Um and we are uh basically soft launching uh solution that we call the take up acceleration engine, which is essentially uh a kind of first layer of mapping of customers' internal data in a way that'll make actioning the take up issues much easier. So understanding where your low-hanging fruit are, uh which you know marketing or commercial approaches have worked, but also figuring out which factors might affect your take-up uh in ways that you don't expect, and sometimes it's external stuff like demographics or um you know various uh elements related to the customers themselves, but sometimes it's also internal stuff like who your subcontractors are, which technology you've deployed, stuff that you wouldn't think has an impact, but actually uh once measured shows that there are impacts. Um and so um yeah, that's uh my big push for the next few months, and um hopefully we'll have stuff that we can show maybe in uh one of our future podcasts.

SPEAKER_01 31:13

Excellent, excellent. Great. And uh I'm um I'll be in the States again next week at ConnectX, which is a sort of Tower Co and 5G and fixed wireless and bit of Wi-Fi uh event. Um, and then um I'll be at the Canadian Telecom Summit the week after in Toronto. Uh so I'm quite curious because I yeah, I sort of dipped in and out of the Canadian telecom market, and it's a bit of a separate world, particularly given sovereignty issues at the moment. So I'll I'll report back on that next time.

SPEAKER_00 31:42

Excellent, excellent. All right, alright. Well, this was fun. Um speak to you soon, Dean. And remember, people, if you like this, uh like, subscribe, share, and uh remember that we're available on all types of platforms in audio and video. And in person and in person as well. And in person, absolutely. Excellent.

SPEAKER_01 32:01

Thank you very much.

SPEAKER_00 32:02

Cheers.

SPEAKER_01 32:02

Bye Dean.