Between roughly 2018 and 2023, Britain allowed more than a hundred rival companies to dig up the same streets and build full-fibre broadband. No other large country attempted competition at this intensity. Cheap money made it possible and a national fibre target made it policy, and for a while it worked: coverage jumped and prices fell. A market this crowded was never going to keep a hundred builders.

My hypothesis is that the market consolidates into three national fibre platforms plus a small tail of niche survivors, with most of the movement complete by 2028. Openreach holds its position through scale and regulation. A CityFibre-led wholesale platform becomes the second pillar by acquisition. Virgin Media O2 and nexfibre form the third. Around them, perhaps six to twelve smaller networks survive on rural economics or unusually disciplined balance sheets. The rest are absorbed.

The evidence runs through the whole report. The sector lost around £1.5bn in 2024 and carries roughly £9bn of debt raised for a rate environment that no longer exists⁹. Average take-up sits near 18% of homes passed, against around 38% at Openreach, and too few paying customers arrived to service the debt². The consolidation itself is no longer hypothetical: the first half of 2026 brought nexfibre's agreement to buy Netomnia for around £2bn¹⁴ and a £2.3bn CityFibre facility raised expressly for acquisitions¹³.

Part Three weighs the outcomes and lands at 68% for the three-platform market, 27% for a slower and messier shake-out, and 5% for fragmented survival. It also states, in advance, the criteria this forecast will be judged against in 2028, and section 3.5 publishes what happened when we stress-tested the whole analysis against its hundred sharpest critics.

The report is written for three audiences, and Part Four carries a briefing for each: the regulator and industry, investors and altnet boards, and every UK household. Each briefing stands alone. Read yours first.

Over the past ten years, more than a hundred companies borrowed billions of pounds to dig up Britain's streets and lay fibre-optic broadband, each racing to reach your home before its rivals. Cheap borrowing made the race possible.

It ended when the cost of money changed. The Bank of England's rate was close to zero when most of these business plans were written; it stands at 3.75% in mid-2026⁵⁶. Debts that once looked manageable no longer are. The largest builders lost £1.5bn between them in 2024 and now owe around £9bn⁹. Customers compounded the problem: on average, only about 18 in every 100 homes that can get these new networks actually buy from them², and a fibre build does not pay for itself at that level.

So bigger companies are buying smaller ones, and the weakest are going bust. A market of more than a hundred builders is on its way to becoming a handful. This report calls that process the Great Consolidation, and expects it to settle, most likely, into roughly three national networks plus a few specialists by the mid-2030s.

An altnet, short for alternative network, is a company building its own full-fibre infrastructure in competition with Openreach rather than renting its lines. At the peak around 2023 there were more than one hundred of them, collectively committing tens of billions of pounds². The conditions were close to perfect: interest rates near zero, a national gigabit target with subsidy behind it, and a regulator that deliberately encouraged rival builds. The result was the fastest infrastructure build in modern British telecoms. Altnet coverage went from a rounding error to 19.7 million premises passed by the end of 2025, up 20% in that year alone². Full fibre as a whole reached 78% of UK homes by mid 2025 and 85% of UK premises by July 2026, and around a quarter of all full-fibre connections are now delivered by altnets^3,5,74.

To follow any deal in this market, one idea is essential: broadband is built in three stacked layers. The physical network is the fibre in the ground, hugely expensive to build and the asset that survives even when a company fails. Wholesale access rents space on that network to retail brands on equal terms; Openreach and CityFibre operate here. The retail brand is the name on the bill, selling the service and handling support. A vertically integrated altnet owns all three layers. A wholesale-only network such as CityFibre owns the bottom two and lets more than 35 retail brands sell on top¹³.

The layers explain why one deal is often really two or three. When nexfibre agreed to buy Netomnia in 2026, the physical network went one way while the YouFibre and brsk retail brands were sold to Virgin Media O2 for £150m^14,48. Reading any acquisition headline, the first question is which layer changed hands.

Three kinds of network now cover Britain, and they overlap heavily. Openreach, the BT Group access division, had passed around 22.1 million premises with full fibre by early 2026 and remains the wholesale backbone most retail providers rely on^3,4. Virgin Media O2, with its build vehicle nexfibre, a joint venture backed by InfraVia, Liberty Global and Telefonica, covers around 8.3 million premises between upgraded cable and new fibre^4,14. The altnet sector is the third group: more than a hundred challengers led by CityFibre at 4.7 million premises at the end of 2025 and Netomnia at 3.0 million, with a long tail behind them^2,73.

These groups do not divide the country between them. They overbuild each other. By early 2026, 13.4 million premises could order from two or more fibre networks, and around 2.5 million from three or more^2,5. The UK now averages roughly two and a half fibre networks per connected household, and three or four in the most contested cities⁷. No other large country overbuilt at this intensity³⁷. By early July 2026, industry tracking put Openreach past 23 million premises and the combined Virgin Media O2 and nexfibre fibre footprint near 8.7 million, so the overlap figures here, measured earlier in the year, are conservative rather than generous⁷⁴.

That overbuild is the fact that makes consolidation arithmetic rather than opinion. When three companies have each spent hundreds of pounds per premises to pass the same house, only one of them usually wins the customer, and none can earn back the dig three times over.

Consolidation is a change of ownership. The cables, cabinets and connections stay where they are, and Ofcom's continuity rules and the buyer's commercial interest point the same way: keep customers connected and billing. Some networks will not consolidate at all. Around 5.4 million premises have exclusive altnet fibre with no Openreach or Virgin Media alternative, and altnets reached 4.4 million premises in the harder Area 3 geography². Rural and community specialists with low debt, loyal customers and no rival on the street are the likeliest standalone survivors. Scarcity protects them.

For the rest, three engines are pulling the market together. CityFibre, the largest altnet and wholesale-only, raised £2.3bn in July 2025 including an £800m facility earmarked for acquisitions, and its chief executive speaks openly of consolidating the altnet sector into a sustainable third national network¹³. nexfibre agreed in February 2026 to acquire Netomnia for around £2bn, creating a platform of roughly 8 million premises and a credible second challenger, a deal now under a fast-tracked CMA Phase 2 review with a verdict due by 15 December 2026^14,57. And Openreach exerts gravity without buying anyone: its scale, reach and wholesale pricing offers such as Equinox pull retail demand toward it, starving thinly used altnets of the customers they need^20,35.

Two active buyers and one gravitational incumbent is the configuration that produces a three-platform market. The 68% central case in Part Three is these engines running to their logical end.

Several altnets now report positive adjusted EBITDA. The same companies are losing large sums. Both are true at once, and the gap between them explains most of this sector. EBITDA excludes interest, depreciation and the capital spent digging, which for a fibre builder are the three largest costs in the business. CityFibre's filed accounts to 31 December 2024 show roughly £6m of adjusted EBITDA becoming a statutory loss before tax of around £360m, with finance costs of about £279m and depreciation near £100m^9,25.

Enders Analysis calculates that altnets need EBITDA margins above 35% simply to reach cashflow break-even, before interest⁹. Few are there. A "profitable" marker in this report therefore means the day-to-day operation covers its running costs, a real milestone for a business still building, and no more than that. Adjusted EBITDA is also non-standardised and differs between companies, so figures are directional rather than strictly comparable³².

The build thesis took hold in 2017 and 2018: near-zero rates, a national gigabit ambition and a pro-competition regulator made fibre cheap to finance, and CityFibre secured Vodafone as a wholesale anchor in November 2017²⁷. Capital flooded in through 2019 and 2020, when CityFibre bought FibreNation from TalkTalk for £206m and lifted its target from 5 to 8 million premises²⁷. Project Gigabit launched in March 2021 with £5bn of subsidy for hard-to-reach premises⁴³.

The turn came in 2022. Altnet numbers passed one hundred just as rates began rising sharply, and sector losses reached £0.76bn⁹. The first cracks showed in 2023: nexfibre acquired the distressed builder Upp after a government-forced sale on national security grounds, a network that had passed 175,000 premises but signed only around 4,000 customers¹⁶. Losses for 2023 reached £1.30bn⁹, and the analyst conversation shifted from build pace to survival.

The shake-out proper began in 2024. CityFibre acquired Lit Fibre in May. Netomnia and brsk combined under the Substantial Group umbrella⁴⁸. Sky agreed in August to use CityFibre's network, roughly doubling its addressable retail market, a deal Enders described as putting CityFibre firmly in the driving seat for consolidation^10,11. Sector losses for 2024 reached £1.5bn⁹. In 2025, FullFibre and Zzoomm completed their merger¹⁹, CityFibre acquired the Connexin network, AllPoints Fibre took shape, and in July CityFibre raised its £2.3bn¹³. Then early 2026 delivered the distress and the flagship deal together: G.Network passed through administration and emerged debt-free¹⁷, nexfibre agreed to buy Netomnia for around £2bn¹⁴, Freedom Fibre and Truespeed agreed to merge, and Gigaclear's lenders took control^18,50.

Since 2020, more than £17.4bn has been invested in UK altnets, over £1,000 for every premises passed, up from under £400 in 2022 as the easy dense areas were exhausted². During the regulatory debates of the period the sector cited more than £25bn committed to BT's competitors³⁰. Most of that capital came as debt, and Enders Analysis estimated in November 2025 that the sector carried around £9bn of it, with financing costs near 121% of revenue, an interest burden it described as looking unpayable under any reasonable scenario⁹. Many altnets owe more in interest than they earn in total.

The losses confirm the direction: £0.76bn in 2022, £1.30bn in 2023, around £1.5bn in 2024⁹. This is a sector whose annual losses widened as cheap-money debt came due in an expensive-money world. The financing wall makes the timetable concrete. Around 47% of UK fibre companies faced refinancing a material part of their debt or equity by 2026, on the 2025 assessment, and the lenders who would normally roll that debt, NatWest and Lloyds among them, have pulled back^21,22. A company that can neither refinance nor service its debt sells, merges or hands the keys to its lenders.

If the financing wall is the trigger, the take-up gap is the underlying condition. Across the sector, average take-up reached 18% of premises passed by the end of 2025, up two points on the year but less than half the roughly 38% Openreach achieves on its own fibre². For every home an altnet connects and bills, it has paid to pass four or five that generate nothing. The debt was raised against all the homes passed; the revenue comes only from the connected.

The unit economics follow directly. The average altnet carries over £500 of net debt per home passed, within a range of £263 to more than £1,000, but over £4,000 per home actually connected²². Enders' verdict is blunt: it is difficult to see a scenario in which retail altnets generate cash returns, even before interest costs⁹. AlixPartners frames the survival side: payback on customer acquisition and connection runs more than five years, and the economics are challenging even if debts are fully written off²².

One number, though, hides a spectrum, and the operators are right to say so. The blended 18% mixes footprints built last year with networks a decade old, and the range across operators runs from roughly 6% to 36%. Exhibit 7 shows it, derived from the filed accounts in Exhibit 18: distress cases sit in single digits, the scaled survivors in the mid-thirties, and maturity clearly matters. The refinement changes the diagnosis for individual networks; it does not change the arithmetic for the sector, because the financing wall prices today's blended number, not the potential of 2030^9,22.

Every entry is a buyer getting bigger, two sellers combining, or a weak network changing hands under pressure. None points toward a more fragmented market. The distress cases repay closer reading. G.Network carried around £300m of debt against roughly 420,000 premises passed but only about 25,000 customers, a take-up near 6%; it was sold to FitzWalter Capital, passed through administration and emerged debt-free in March 2026 with the network operating throughout¹⁷. At Gigaclear, eleven lenders including the taxpayer-backed National Wealth Fund, NatWest, Lloyds, HSBC and ABN AMRO took control from equity owners Infracapital, Equitix and Railpen, accepting a haircut of up to around 40% on near £1bn of debt^18,50.

Exhibit 9 puts the crystallised pain in one picture. Roughly £700m of debt value has already been written off or extinguished across just the two 2026 distress cases, before counting the equity wiped out on top, which is undisclosed. The pain fell on investors and lenders while both networks were preserved and re-homed. This is what distressed consolidation looks like in a sector built on real, working assets, and it is why even the slower 27% scenario is not a consumer catastrophe.

Plotting who is exposed against how badly, most consolidation risk is financial and lands on the investors and lenders who chose it. Equity write-downs and lender haircuts sit exactly where capital markets place them. Staff and suppliers carry real but bounded exposure as construction gives way to connection.

The cell that deserves genuine public attention is the household relying on a single altnet with no Openreach or Virgin alternative, especially the roughly 940,000 hard-to-reach premises covered by Project Gigabit contracts⁴³. If one of those networks fails badly, its customers have no easy fallback, and Part Four returns to what that means for the regulator and for the households themselves.

Customer experience data adds the human dimension. Small challengers dominate satisfaction measures: Gigaclear and Zzoomm score 4.7 on Trustpilot against BT at 1.3 and Virgin Media at 1.5, with self-selection bias applying⁴⁶, while Zen topped the Which? survey of 5,235 members at 84% with Hyperoptic and Community Fibre close behind, and Virgin Media, TalkTalk and NOW Broadband at the bottom on 59%, 59% and 54%⁴⁴. Ofcom's complaints data runs the same direction, though Virgin Media's turnaround from the worst record in Q4 2023 to among the best by Q4 2025 shows service reputations can be rebuilt at scale^45,55. The repair is recent, though: on 8 July 2026 Ofcom fined Virgin Media £28 million, its largest ever penalty under its consumer protection rules, for obstructing customer cancellations between January 2022 and September 2024⁷⁵. The challenge for the surviving platforms is keeping challenger-level service at incumbent scale, because churn is the enemy of a debt-laden network: a provider that cannot keep customers cannot pay back the cost of connecting them.

TalkTalk is the cautionary tale on the other side. A 2015 cyber attack accessed 156,959 customer accounts, cost around £60m, drew a then-record £400,000 ICO fine and lost the company 101,000 customers³⁴. From 4.1 million customers after the Tiscali acquisition, the base fell to about 3.2 million by 2025, with roughly 400,000 lost in 2024 alone, while the 2021 take-private left leverage near fifteen times earnings on about £1.2bn of debt³⁶. A 2024 refinancing, a 2025 split into consumer, business and wholesale arms and a £100m injection pushed maturities to 2028³⁶. Heavy debt plus lost customer trust produces a long, grinding series of refinancings rather than a clean recovery, and it is the outcome the weakest altnets may not avoid.

How the scheme worked. The Gigabit Broadband Voucher Scheme launched nationally in 2018, and its first £67m phase was fully used within two years⁶⁴. A second, rural-focused wave followed from May 2019, and the scheme was folded into Project Gigabit from 2021 with a £210m allocation inside the £5bn programme⁶⁴. The mechanics are supplier-led: homes and businesses in eligible areas pledge vouchers now worth up to £4,500 per premises, a registered builder pools them into a project, builds the connection, and is paid by Building Digital UK only once each connection is live, with local authority top-ups sometimes adding thousands more⁶⁴. By April 2024 around 118,000 vouchers had been used to fund a connection⁶³.

What it built. Vouchers made tiny rural builds bankable one cluster at a time, and they part-created the niche tail this report forecasts as survivors. The exemplar is B4RN, Broadband for the Rural North: a community benefit society and registered voucher supplier that has passed around 32,000 of some of Britain's hardest premises, often with volunteer labour, and which is structurally acquisition-proof because its members own it⁶⁶. The voucher model and the community model fit together: pooled public money plus pooled local effort reaches the places no commercial case could.

What changed. Procurement has superseded vouchers as the main channel for public money. As Project Gigabit's contracted builds scaled, the vouchers' share of subsidised delivery fell from around 37% to around 16%^63,64, and voucher availability was progressively paused or withdrawn in areas earmarked for contracts, a mechanism that has repeatedly stranded community groups mid-application⁶⁴. The end is now dated: BDUK announced in early 2024 that the scheme would close at the end of March 2028, with the final vouchers issued by 31 March 2027, and adjusted the closure mechanics in June 2026⁶⁴. Every voucher-dependent business model in the country now has a countdown running.

Where it went wrong, and where the work goes. Broadway Partners, a voucher-era rural builder, entered administration in 2023 and its assets were absorbed by Voneus⁶⁷. And when FullFibre, Freedom Fibre and Voneus stepped back from Project Gigabit lots through 2025 and 2026, the descoped premises were re-let to Openreach under the cross-regional Type C framework worth around £1.2bn⁶⁸. The pattern deserves plain statement: when subsidised challengers stumble, the subsidised work flows to the incumbent. Parliament's Public Accounts Committee has pressed the value-for-money question throughout, and continuity of subsidised contracts through ownership changes is now a live procurement issue rather than a hypothetical one^65,68.

The old analogue phone network switches off on 31 January 2027, a date Openreach says will not move again, with around 1.9 million lines still to migrate as of June 2026, down from more than 16 million when the programme began⁶⁹. The forcing is now financial as well as calendar: legacy line rental prices rise 20% from 1 April 2026, a further 40% from 1 July and 40% again from 1 October, roughly doubling across the year⁶⁹. Behind the switch-off sits the longer Exchange Exit programme: Openreach plans to shrink from around 5,600 exchanges to roughly 1,000, with about 105 priority closures by 2030 and the majority of the remaining 4,500 exiting through the early 2030s⁷⁰. More than 1,000 exchanges covering roughly 8.9 million premises were already under stop-sell by mid-2026⁶⁹. The ageing network it retires still consumes around 1% of Britain's energy⁶⁹.

The consolidation reading cuts both ways, and lands one way. Forced migration is a demand tailwind: nearly two million lines must land on modern networks within months, and some land on altnets. But exchange exit is also a new cost line for every altnet with kit in a closing building: operator estimates for relocation and re-routing run from £100,000 to £200,000 per exchange at Netomnia to £1.4m per exchange at Neos Networks⁷¹. Scale absorbs a relocation programme across three-year notice periods; sub-scale does not. The switch-off accelerates the sorting.

The starting point is not opinion but reference class: how did comparable situations actually resolve? Britain has run this experiment before. Cable television was built by more than 130 franchises in the early 1990s; after heavy debts and two restructurings they collapsed into two operators, then one⁵¹. Mobile went from many operators to a handful, most recently through the Vodafone and Three merger³¹. The base rate for capital-heavy, debt-funded networks is that they consolidate toward a very small number.

Two adjustments then apply. Pushing toward more consolidation than the base rate: the UK overbuilt harder than the cable era, carries around £9bn of debt, and converts only 18% of homes passed^2,9. Pushing toward less, or slower: some networks carry little debt, community models cannot easily be bought⁶⁶, and a regulator alert to duopoly risk could block deals.

Three approaches triangulate the result. Structural extrapolation follows the balance sheets: which networks can plausibly service or refinance their borrowing, given the 2026 to 2027 refinancing wall²². Reference-class forecasting weighs the international record set out in section 2.7. Scenario weighting then combines both with the structural facts, two active buyers, one gravitational incumbent and a single genuine wildcard in the CMA, into explicit probabilities. The result, held since first analysis and re-tested against everything in Part Two: 68%, 27%, 5%, published 12 June 2026.

The central case runs in three phases: consolidation to 2028 as the largest deals close and the refinancing wall forces exits, platform completion to 2031 as three national wholesale platforms take shape, and maturity to 2035 as take-up climbs and pricing power returns.

Movement between scenarios is specified in advance. The CMA clearing nexfibre and Netomnia at Phase 2 largely unconditionally, verdict due by 15 December 2026⁵⁷, two or more further large deals closing in 2026, take-up staying below 20% and rates staying elevated would push scenario A toward 80%. A block or heavy conditions, a material fall in rates, or take-up accelerating past 30% in maturing footprints would lift scenario B sharply, and any material change to these probabilities will be published at broadbandswitch.uk with reasoning attached.

The best evidence-based case that this forecast is wrong runs as follows. Predicted altnet consolidation has been imminent for years and has repeatedly arrived more slowly than forecast, a point industry observers make directly²⁸. Infrastructure funds are patient capital and can hold loss-making networks rather than sell at a loss. The CMA, having scrutinised Vodafone and Three closely, may block or heavily condition the largest fibre deals: CityFibre itself argues that nexfibre buying Netomnia would re-establish an ineffective duopoly, citing an 80% overlap between the two networks⁴⁹, and Point Topic puts the overlapping footprint at around 832,000 premises⁵⁸. The deal now sits in a Phase 2 investigation, albeit one the parties themselves asked to be fast-tracked into, with a verdict due by 15 December 2026⁵⁷. And if rates fall materially, the refinancing wall softens⁵⁶. On this view the answer is a long, untidy decade of survival rather than three platforms by 2028.

Each point is real, and the 27% weight on the slower scenario exists because of them; it is a substantial probability rather than a token hedge. None, however, changes the structural facts. Patient capital still needs a return eventually, and financing costs of 121% of revenue do not wait a decade⁹. Even a blocked nexfibre deal likely redirects Netomnia toward CityFibre rather than preserving it standalone, so the platform count still falls²⁴. The strongest exhibits in the counter-case, the CMA and the rate cycle, are arguments about speed, not destination.

Finding one: the blended take-up critique. The operators' best punch lands squarely on the 18%: a blended average mixes footprints built last year with networks a decade old, and mature footprints convert far better than the average implies. The critique is fair, and this edition answers it with evidence rather than defence: Exhibit 7 now shows the spectrum by operator, from roughly 6% at the distressed end to the mid-thirties at the scaled survivors, derived from the filed accounts²⁵. The refinement matters for diagnosing individual networks. It does not rescue the sector arithmetic, because the refinancing wall prices today's blended cash flows, not the take-up a footprint might earn by 2030^9,22.

Finding two: the ageing-figure audit. Forecasts rot quietly, through figures that were true when written. The audit flagged the 47% refinancing-by-2026 finding as a 2025-vintage assessment, now labelled as such wherever it appears²², and drove the re-verification of the fastest-moving inputs for this edition: the Bank Rate hold of 18 June 2026⁵⁶, the CMA Phase 2 fast-track of 1 July 2026⁵⁷, the voucher scheme closure timetable⁶⁴ and the copper switch-off status⁶⁹.

Finding three: forecast accountability. The forecasting community's standard is not being right, it is being auditable: publish the probabilities, log every revision with reasons, score in public⁷². This report now commits to exactly that. The 12 June 2026 probabilities stand on the record, every material revision will be published at broadbandswitch.uk with reasoning attached, and the 2028 scorecard lands whichever way the count goes.

The settlement expects consolidation. On 17 March 2026 Ofcom published the final statement of its Telecoms Access Review for 2026 to 2031^30,54. Openreach retains significant market power across the markets that matter and must keep offering regulated access to its ducts and poles, with the fair bet principle retained: Openreach keeps the upside on its fibre investment in exchange for having carried the risk³⁰. The structure that built a hundred networks is deliberately left standing. What has changed is the assumption underneath it: the review accepts that the market is more fragmented than anyone anticipated and that this is not a stable end state, while the government's draft Statement of Strategic Priorities says outright that the fixed market will consolidate over the review period³⁰. When the rule-setter and the policy-setter both write consolidation into their forward plan, it has stopped being a forecast and become an assumption. Ofcom indicated it would welcome a CMA review of the nexfibre and Netomnia deal, and on 1 July 2026 that review was fast-tracked to Phase 2 with Ofcom engaged throughout⁵⁷. The settlement permits the wave; the CMA polices its shape. The competitive geography is explicit: Area 2, where rival build is present or expected, holds around 54% of postcode sectors for leased lines, and Area 3, where competition is limited, around 34%³⁰.

The measurement problem. Almost every argument about altnets is really an argument about definitions. Premises passed measures construction: the sector reached 19.7 million on this measure by the end of 2025². Ready for service is narrower and inconsistently reported. Live connections is the number that pays the bills: 3.5 million, producing the 18% take-up rate at the centre of this report². A network is financed against premises passed and survives on live connections. For a merger review the choice of measure is decisive: counted by premises passed, two overlapping networks look like vigorous rivalry; counted by live connections, the same streets may show one network with customers and one nearly empty.

A single disclosure would close most of this register: audited take-up and net debt on a common definition, published even annually. Tracking the wave honestly takes five indicators read together: the count of independent networks above roughly 250,000 premises, the take-up trajectory, deal cadence, distress events, and the overbuild ratio, whose peak marks the turning point of the cycle. Premises passed alone flatters the market; distress events alone catastrophise it.

The wider cast. Consolidation moves far more parties than the three platforms. The acute exposures are Project Gigabit, where subsidised contracts cover roughly 940,000 hard-to-reach premises and descoped lots have already flowed to the incumbent's £1.2bn framework^43,68, and the pension and infrastructure funds, including a taxpayer-backed lender at Gigaclear, that carry the write-downs⁵⁰. Wholesale-only retailers can be stranded if their host network changes hands, contractors feel the build-to-connect shift well before the deals make news, and councils face fewer, larger counterparties. The deal is signed by three parties and its consequences are shared by many.

Three or more weak answers place the network on the consolidated side of the line, where the realistic choices are sell, merge or restructure, and the best outcomes go to boards that move early on their own terms. Five or six strong answers describe a consolidator, or a genuine standalone survivor.

The line itself. The consolidators are well-funded, lower-leverage platforms with the equity and the appetite to buy: CityFibre after its £2.3bn refinancing¹³, nexfibre after its move on Netomnia¹⁴. The consolidated passed homes faster than they connected them and borrowed to do it; G.Network's administration and Gigaclear's lender takeover are this side of the line in its harder forms^17,50. A third destination exists narrowly: rural-exclusive or community-owned networks with low leverage and loyal, high-take-up customers, which survive precisely because they never played the overbuild game⁶⁶. The deciding variable is debt-repayment capacity. Strip away the branding and the question is arithmetic: can take-up multiplied by revenue per customer service what was borrowed to build? Where it can, the network has options. Where it cannot, the network has a timetable set by whoever holds the debt.

The refinancing wall. Around 47% of UK fibre companies faced refinancing a material part of their debt or equity by 2026, on the 2025 assessment, with financing costs across the sector near 121% of revenue and the traditional lenders pulled back^9,21,22. Debt raised at low rates must be replaced at higher ones at the precise moment take-up has yet to catch up with the build, and every month of delay raises the price of the eventual fix.

Wholesale pivot or retail contest. Faced with the take-up problem, networks have taken two roads. The wholesale pivot sells capacity to established retail brands: CityFibre is the purest case, wholesale-only and carrying Sky, Vodafone and TalkTalk^10,13, with aggregation platforms now bundling several altnets into a single shop front. A wholesale network's value is its reach and the quality of its fibre, assets that survive a change of owner and combine cleanly with a buyer's footprint. The retail contest pits a sub-scale altnet against BT, Sky and VMO2 for every customer, with payback on acquisition costs already running past five years²². A retail base can melt away when a brand disappears, which is why retail customers are often the least durable thing a distressed altnet owns. For value, a clean wholesale network with reach tends to command a better outcome than a retail book of uncertain loyalty.

The distress playbook. A managed exit runs in four stages. The squeeze: take-up lags, interest rises, the next round arrives late or on punitive terms, covenants tighten. The search: a sale or merger sought from relative strength, where the best outcomes live. The lender reset: at Gigaclear, eleven lenders including a taxpayer-backed fund took the keys, accepting a haircut of up to roughly 40% on near £1bn of debt, with the equity wiped^18,50. Administration and rebirth: G.Network filed in January 2026 and emerged weeks later debt-free under new ownership, with the network and customers carried through¹⁷. Both 2026 cases teach the same timing lesson. The expensive mistake is waiting until the lender, rather than the board, decides the timetable.

If your provider is bought. The connection is the asset the buyer paid for, so the overwhelming likelihood is that your broadband keeps working, often with no interruption at all. What can change is the brand on the bill, the account system and, over time, the price: the most common real-world effect of consolidation arrives at your next renewal, when an introductory rate can step up. What can improve is service: a larger, better-funded owner can bring stronger support and faster fault repair than a stretched start-up. Before any transfer completes, record your current price, speed and contract end date; your existing terms should be honoured, and if a new owner changes them to your material disadvantage mid-contract you typically gain the right to leave without penalty.

If your provider fails. Even in administration or a lender takeover, the fibre is the most valuable thing the company owns, and administrators and new owners keep it running because a dark network is worth far less than a lit one. G.Network's customers were carried through its 2026 administration; Gigaclear's lender takeover was a balance-sheet event its customers mostly never noticed^17,50. The genuine risk case is a small, exclusive network with no Openreach or Virgin alternative that is wound down rather than rescued. It is uncommon, because exclusive rural networks are often the most attractive to a rescuer, but it is the case worth knowing your position on in advance, and it is where a satellite service can serve as a genuine stop-gap.

The gap households should know about. Unlike energy and water, broadband has no statutory safety net if a network operator fails. When the energy supplier Bulb collapsed in 2021, the law let the government step in, at an eventual cost of £6.5bn⁵⁹. No equivalent law exists for broadband: households are protected by the value of the fibre and by ordinary insolvency rules, not by a guarantee. For an essential service, that is a policy gap for government to close rather than a reason for alarm, and this report's regulatory briefing places it on the record.

The choice window. In the millions of homes passed by two or three networks, competition is keeping prices keen and switching generous^2,3, and that pressure eases as the number of rivals on a street falls. Industry rules make moving straightforward, and in most cases the new provider manages the switch. Sometimes the right move is none at all: a strong connection at a fair price is not worth abandoning because the owner changed.

Investment totals depend on scope, so this report shows them separately rather than combining them. Assembly Research, commissioned by nexfibre, reports £21.4bn invested across the whole sector from 2020 to 2025; Intelligens Consulting tracked top-10 altnet secured investment across 2022 to 2024; and on the wider private-equity-backed definition covering 2018 to 2023, more than £40bn entered UK fibre^7,14,47. The money coming in has collapsed either way: new debt financing fell from £1.2bn in 2023 to £600m in 2024 and about £170m by early 2025, across a handful of deals²².

Fibre is valued per home passed, and the nexfibre and Netomnia deal shows why headline prices and cash reality diverge. The average altnet carries £263 to more than £1,000 of net debt per home passed²². The £2bn ask equated to roughly £700 to £900 per home. Enders' Karen Egan set the cash ceiling near £500, because new fibre can be built for about that and a cable overlay added for £100 more, and put a fair overlap-adjusted price below £300 a home given that Netomnia and Virgin Media O2 overlap heavily⁹. Wider context: scaled wholesale fibre trades at 8 to 12 times EV/EBITDA, with 15 to 20 times reserved for strategic scarcity⁶², and build costs run from £303 per premises in urban footprints to £1,803 in deep rural⁶¹.

The funding markets consolidated before the operators did. Analysys Mason found that since 2021 about 70% of capital raised by challenger fibre operators went to three firms: CityFibre around 60%, roughly £6bn, Netomnia around 7% and Gigaclear around 4%, with the remaining 29% spread across every other altnet. Over the same window the number of mergers and acquisitions matched the entire previous decade and their combined value surpassed it by around three and a half times⁶⁰. Investors stopped backing the field and backed a shortlist.

Behind the brands sit a small number of infrastructure funds, several holding more than one altnet, and those investors are the real consolidation vectors. DigitalBridge backs both Netomnia and Freedom Fibre. The National Wealth Fund, the taxpayer's vehicle, now sits among Gigaclear's owners and CityFibre's lenders⁵⁰. When one investor holds several altnets, merging them is the natural next step, which is what Fern Trading did with its four networks. The clearest read on appetite came in June 2026, when the Financial Times reported KKR and Warburg Pincus exploring exits from Hyperoptic and Community Fibre¹⁵: appetite for retail-facing altnet risk is cooling while appetite for scaled wholesale platforms holds, so the next wave of deals will be investor-led.

The same headline, "altnet in trouble", hides three realities. The going-concern note: an auditor flags material uncertainty, as Airband's 2024 accounts do in stating that funding required from 1 September 2025 "has not been identified and secured", and as Hyperoptic's 2022 and 2023 accounts did before its KKR backing and Openreach wholesale pivot⁸; the outcome is rescue or pivot if a backer steps in. Negative equity: liabilities exceed assets, confirmed at Ogi, WightFibre, Truespeed and Fibrus among others; these operators are not yet failing, but they cannot raise equity at a sensible price, so they become acquisition targets, often to a same-investor sibling. The lender takeover: debt cannot be serviced, so lenders act, as at Gigaclear and, through administration, G.Network^17,50.

The leading indicators are public. When a sponsor declines to fund a committed tranche, or an auditor adds a material-uncertainty paragraph, history suggests roughly two to four quarters to a restructuring. On that test, the operators in deepest negative equity with the largest lender impairments are the names to watch through 2026 and 2027.

Every consolidation principal named in this report was checked against the public register, and any reader can repeat the check at Companies House. The majors: CityFibre Holdings Limited 07488363, Nexfibre Networks Limited 12175177, Netomnia Limited 12008248, Substantial Group Limited 12315611, Gigaclear Limited 07476617, G.Network Communications Limited 10057745, Hyperoptic Ltd 07222543, Community Fibre Limited 07413288. The full validation table, with incorporation dates, registered offices and persons of significant control, carries over from the first edition as an appendix.

Jersey, Guernsey and the Isle of Man sit outside this report's main forecast but inside its logic. Where the mainland built a hundred rival networks, the islands never fragmented: each is served by a single dominant operator, with retail competition layered over one wholesale network. Jersey's States-owned JT completed 100% fibre to the premises years ago, one of the first jurisdictions in the world to do so³⁹. Guernsey's Sure is delivering a £37.5m States-backed rollout, near-complete by the end of 2026, with JT reselling over its wholesale network⁴⁰. On the Isle of Man, Manx Telecom is the incumbent and designated fibre provider, including a £10m contract covering nine hard-to-reach intervention zones under the National Broadband Plan⁴¹; island data is largely operator-sourced and graded accordingly.

Consolidation arrived on the islands in 2026 by a different route. JT Group and the infrastructure investor CVC DIF agreed to acquire Manx Telecom, approved in April 2026 under the National Security and Investment Act 2021 subject to conditions on continuity of services to UK government and the establishment of a cyber-security group, creating the largest full-service telecoms provider across the Crown Dependencies⁴². The logic is the mainland's, scale through combination, reached from a starting point of monopoly rather than overbuild.

The author. Dr Alex J. Martin-Smith CMgr MBA LLM DBA is the founder and director of the company behind BroadbandSwitch.uk and writes its authority reports, of which this is the twenty-seventh, applying a consistent house method: answer first, grade every figure, and never use a statistic without a value, a date, a scope and a named source.

The reviewer. This report was reviewed by Adrian James, who challenges the analysis, the sourcing and the forecast before publication. The review does not soften the conclusions; it tests whether they survive scrutiny. Any errors that remain are the author's own.

About BroadbandSwitch.uk. BroadbandSwitch.uk helps UK households and businesses understand the broadband market and switch with confidence. Alongside its comparison tools, it publishes evidence-graded authority reports on the forces reshaping UK connectivity, from the analogue phone switch-off in the previous report to the fibre market consolidation forecast in this one, designed to be read by households, boards and regulators alike, and to be checkable by all three.

Suggested citation. Martin-Smith, A. J. (2026). The great consolidation: Forecasting the future of the UK ISP and altnet market (Report No. 27). BroadbandSwitch.uk.