UK fibre build and overbuild

thinkbroadband publishes excellent monthly data on fibre broadband deployment in the UK (Sep 2024 figures here). In this note I've taken a look at these figures over time, to see what they tell us about build, overbuild and the pace of deployment.

The first thing to note is the relatively good coverage now available. 28m premises (85%) have Gigabit broadband, and 23m (71%) have full fibre broadband. The difference between these two figures is likely to be primarily 'Virgin solus' premises - areas where the only provider of Gigabit is Virgin Media.

By late 2021, Virgin Media had upgraded its cable network to DOCSIS 3.1, so it could offer Gigabit to the 16m premises then in its footprint. As of January 2022, there were 10m likely 'Virgin solus' households. In other words, for roughly two-thirds of the households in Virgin's 16m footprint, it was the only supplier of Gigabit broadband. However, today, there are just 4.5m 'Virgin solus' premises, or less than one-third of Virgin's total. This points to the erosion of what was once a significant technical advantage for Virgin Media, where in many parts of the country it could offer far higher speeds than its rivals.

thinkbroadband's data also allows us to look at how full fibre coverage splits between the incumbents (Openreach and KCOM) and the altnets (including Virgin Media FTTP), and the degree of overbuild - premises with offers from both an altnet and an incumbent. Note that there will be some additional premises where altnets have overbuilt each other but no incumbent FTTP is available. While it's not possible to identify these from the thinkbroadband figures, the number is likely small.

The chart below shows the number of premises that are OR/KCOM only, altnet only, and the number that are overbuilt (either because - say - OR has built where there was already altnet FTTP, or vice versa):

As of September 2024, there are roughly equal numbers of premises in each category. In other words, we can divide UK fibre coverage into thirds - altnet only, OR/KCOM only, and overlap. This implies that roughly half the OR/KCOM total footprint overlaps with altnet coverage, and vice versa.

It is striking that neither OR/KCOM nor altnets are seeing material growth in their 'solus' premises. This obviously doesn't mean that neither are are deploying fibre to new areas, but rather that as quickly as they add new solus premises, they are being overbuilt in their existing footprint. As a result, the number of overbuilt premises is growing significantly.

Whether this is a good or a bad thing depends on your perspective. For Ofcom, it points to an increasingly competitive ultrafast market, albeit still unevenly distributed. For investors (both in altnets and BT), it is of course more of a concern. The available market is likely to be divided between more players, and the higher quality and speed of fibre is no longer a USP (a critical challenge for future marketing messages).

Looking ahead, the degree of overbuild may create a further challenge for altnets, in that it suggests a growing portion of the country will be served by three fibre providers, strengthening Openreach's case for at least local deregulation.

The next chart (also shown at the top of this article) shows how much of fibre build is overbuilt, both cumulatively in red and monthly in yellow:

As we have noted, roughly a third of FTTP premises (cumulatively) have been overbuilt, and this portion is continuing to rise, since for new build in each month, the figure is around 40%.

One of the reasons for this is of course the diminishing greenfield opportunity. Our next chart shows the build opportunity for altnets. I've made the simplifying assumption that altnets would rather not overbuild other altnets, though certainly this does happen.

As of January 2022, an altnet choosing its next deployment area would have had 21m greenfield premises to choose from, and a further 5m where the only FTTP competition was from Openreach. Today those figures are 9m and 8m respectively. Given that an appreciable portion of those greenfield sites are likely uneconomic for one reason or another, it highlights why so much recent build is in fact overbuild.

This is one of several factors that has made investors increasingly cautious about funding continued deployment, and this is evident in thinkbroadband's figures:

While OR/KCOM have had a (relatively) steady pace of deployment, the combined altnets' build rate peaked in Q3 2023, and has fallen by about 40% since. That said, altnet deployment rates still (in aggregate) match those of OR, and are a vital part of the overall picture.

Finally, we take a quick look at regional figures for full fibre coverage. With a few exceptions, they are surprisingly consistent, currently mostly in the range 65-75%. Northern Ireland is well ahead (in part thanks to Project Stratum), and Yorkshire & Humber is second (likely due to KCOM's early deployment). Scotland lags due to its challenging geography, though not dramatically so.

London is an interesting case however. It has dropped sharply down the rankings over the last year, from 3rd to 6th. Apart from Northern Ireland (which has almost completed FTTP deployment), London has the lowest rate of fibre build in the UK, by some margin. Likely factors are the expense of deployment in London streets; widespread competition from Virgin Media; the challenges of access to certain MDU's; and so on.

In conclusion, thinkbroadband's granular data is entirely consistent with the macro trends in the industry. Rapid deployment in 2023 has given way to a more sober build rate, and overbuild is increasingly widespread. It seems plausible that build rates will fall further because of the diminishing greenfield opportunity, and (an associated point) because certain regions may be approaching the limits of what is commercially viable.