An infrastructure platform wants to own fibre. It can build a network, buy a finished one, or buy a small network and build on it. This is the whole thing worked through at the level a deal team would run it, the three routes compared live, the auction end to end, the capital structure and financing in depth, the valuation maths, and exactly who you call.
Tip, hover (or tap) any underlined term for a definition.
Our platform, a mid-market infrastructure fund with a fresh allocation to digital, has set its thesis: full-fibre is a once-in-a-generation rebuild of the access network, with monopoly-like local economics once a street is built and penetrated. The UK is the worked example: the copper network switches off in January 2027, government backs gigabit coverage through Project Gigabit, and the market is fragmented across more than a hundred altnets, many sub-scale and loss-making (sector accounting losses passed £1.5bn in 2024). That fragmentation, and the overbuild it created, is forcing consolidation. The question is not whether to enter, but how.
Fibre has the infrastructure characteristics a fund underwrites, an essential service, very high barriers to a second build, 30-year-plus asset life, and a recurring, inflation-linked subscription once connected. But it is not a regulated annuity on day one: a greenfield network is a development asset that burns cash building homes passed, then inflects to a high-margin annuity as penetration climbs. Three variables decide the case.
The single biggest risk. If a second or third network builds the same streets, penetration and ARPU both fall. Diligence orbits one question: is this footprint defensible? A network that is sole-fibre in its towns is worth a different multiple to one fighting two rivals.
The flywheel. Build cost is fixed per home passed; the return is earned on take-up × ARPU. Below ~25–35% a network loses money; at 45%+ it is a star. The penetration ramp by cohort, how fast each year’s build fills up, is the most-debated line in the model.
Front-loaded capex of £300–£600 per home passed in efficient urban builds (much more rural), before the cash turns. You need a build machine, civils, wayleaves, vendors, and committed capital to fund the J-curve without stalling mid-build.
The mandate sets the frame: a core-plus / value-add return target (low-to-mid-teens equity IRR), a 7–12 year hold, a wholesale / open-access model to stay capital-light on retail, and a footprint we can scale to an exit that a larger infrastructure fund, a strategic, or the public markets will buy. The reference build for how this is financed is CityFibre’s 2022 package, c.£4.9bn of debt plus equity from Antin, Goldman Sachs, Mubadala and others, to fund a rollout to ~8m premises: the largest full-fibre financing Europe had seen, and the template a new entrant is measured against.
Three ways in, the same end-state footprint. They differ in who creates the value, you, at cost, taking the risk, or a seller you pay a multiple, how fast you reach scale, how much equity is at risk through the J-curve, and how much execution rests on your own team.
| Dimension | Build it yourself | Buy a full network | Buy small & build |
|---|---|---|---|
| Speed to scale | Slow, years of build | Instant, at completion | Medium, platform now, build on |
| Value capture | Highest, built at cost | Lowest, pay a full multiple | High, blend |
| Execution / build risk | Highest | Lowest, already built | High, but a team in place |
| Penetration risk | Full, ramp from zero | Low, proven take-up | Mixed |
| Equity at risk | Phased, lower total | Large, day one | Phased |
| Competition for the deal | None, you originate it | High, auctioned | Medium |
| Entry valuation basis | build cost / home | EV/EBITDA (full multiple) | EV/home + build at cost |
| Best when… | defensible whitespace & a build machine | scale fast, low risk, can win the auction | a credible sub-scale platform is for sale |
Now put numbers on it. The comparator runs one funding model three ways over the same target footprint, a cohort penetration ramp, a capex/acquisition funding waterfall (senior debt to a cap, equity fills the gap, surplus sweeps debt then distributes), and an exit at a market multiple.
—
—
Each cell re-runs the greenfield route at that penetration and exit, holding everything else at the current settings; the outlined cell is today’s case. Penetration is the dominant axis, the visual case for why the overbuild map matters more than the financing.
How it works. All three routes converge on the same end-state footprint, so they are compared like-for-like. Each year, homes passed are built on a schedule (or acquired at entry); penetration ramps each cohort to the steady-state level; revenue = connected homes × ARPU (growing with inflation) × 12; EBITDA = revenue − opex. A funding waterfall draws senior debt up to the cap (a multiple of mature EBITDA), equity fills any gap, and once cash-generative the surplus sweeps debt then distributes; the exit adds the equity value at the chosen multiple. Building flatters the point estimate because it books value at construction cost, but it carries the execution and penetration risk a single IRR can’t show. Illustrative only; not a forecast or investment advice.
Everything reduces to four numbers and one curve. Get fluent in these and the diligence, the model and the negotiation all follow.
Homes passed is the footprint, premises the fibre runs past, ready to connect. Homes connected is the subset actually taking service. The ratio is penetration. You spend to pass; you earn on connect.
Build cost per home passed (civils, fibre, PON electronics, ~£300–£600 urban, far higher rural) is the fixed bet. Connection cost (the drop and the ONT, ~£150–£300) is incurred only when a customer signs, capex that follows revenue. Using the incumbent’s ducts and poles via PIA is the main lever on build cost.
ARPU is monthly revenue per connection. A wholesale network charges ISPs a rental (~£10–£25); an integrated one bills the consumer (higher ARPU, but retail cost and churn). Most infra capital prefers wholesale, capital-light, neutral, contracted via an anchor tenant.
Because cost is fixed per passed home and revenue scales with take-up, early years are cash-negative, the bottom of the J-curve. Past break-even penetration the network inflects into a high-margin annuity. There are two break-evens to track: EBITDA break-even and, later, cash break-even (after interest and capex). The art is funding the trough without a covenant breach.
How you value a fibre network changes with maturity, and the gap between the two methods is the whole game:
| Stage | Primary metric | Why | Typical range |
|---|---|---|---|
| Early / building | EV per home passed + a DCF of the ramp | EBITDA is negative or meaningless; you are buying a footprint and a forecast | ~£200–£900 / home, by penetration & overbuild |
| Mature / penetrated | EV / EBITDA | A stable, cash-generative annuity is valued like one | low-to-high teens × |
The builder’s edge is the build multiple. If you can build and connect a home for, say, £500 of passed cost and that home (at 42% penetration, £17 ARPU) contributes EBITDA worth ~£60 at a 14× exit… the arithmetic is the other way round: the network is created at an implied entry multiple of roughly build cost ÷ mature EBITDA per home, often 8–11× if the build is efficient, and exits on a mature 13–15×. Buying a finished network hands that spread to the seller; building keeps it (if you execute). That single fact is why the build and buy-and-build routes out-return buying in the comparator, and why the return is earned, not bought.
The valuation does not happen in a vacuum, you triangulate against where comparable digital-infrastructure assets have actually traded. A selection (announcement values, as reported):
| Target | Acquirer | Year | Value | Read-across |
|---|---|---|---|---|
| Netomnia / Substantial | nexfibre (InfraVia, Liberty, Telefónica) | 2024 | EV €1.5bn | UK altnet consolidation; reported ~10× forward EBITDAaL |
| Telecom Italia NetCo | KKR-led consortium | 2024 | EV ≤€22bn | National fixed network carved out as pure infrastructure |
| CityFibre (financing) | Antin, Goldman, Mubadala + | 2022 | £4.9bn debt | The UK altnet build template, scale funds the J-curve |
| CyrusOne | KKR + GIP | 2021 | $15bn | Data-centre take-private, digital firmly an infra asset |
| AirTrunk | Blackstone + CPP | 2024 | A$24bn | Hyperscale at scale; the AI-demand bid for capacity |
→ See the full digital-infrastructure deal set in the deals database
This follows the eight phases of the M&A lifecycle, specialised to fibre. Building is an internal capital-allocation and financing exercise; buying is a competitive two-round auction; a buy-and-build runs the full process for the platform, then a lighter version for each bolt-on. Indicative timing is for a buy.
Fix the thesis precisely: geography, footprint, wholesale vs integrated, return and hold, and the overbuild map, which postcodes are unbuilt, who else is building, where you can be first or second. Source the entry: for buy/B&B, mandate a buy-side adviser and work the altnet founders, corporate carve-outs (a telco spinning out infra) and secondaries; for build, line up the build platform, land/wayleave access and an anchor ISP.
In real life: brief 2–3 infrastructure-focused advisers, build your own overbuild map from regulator and commercial data, and decide your route before you see a teaser.
A teaser arrives; sign the NDA and receive the Information Memorandum, vendor due diligence and access to the data room. For fibre the case lives in a few exhibits: the homes-passed build plan and actuals, the penetration curve by cohort, ARPU and churn, build cost per premise (planned vs incurred), the ISP/wholesale contracts, and the overbuild exposure. Build a first-cut model and a value range.
In real life: the seller runs it to a process letter; some sellers attach a stapled financing to set a debt benchmark. Your output is a fast internal screening paper.
Submit the NBO: a value range (on EV/home for an early-stage network, EV/EBITDA for a mature one), structure, financing approach, conditions and timetable. Win a place in round two and the deeper data room. For a build, the equivalent gate is an Investment Committee mandate to commit development capital.
In real life: the NBO is a relationship document too, sellers shortlist on price and certainty (your track record of closing and funding the build).
The core work, run as parallel workstreams: Commercial, the overbuild and footprint analysis, penetration-curve realism, ARPU sustainability and ISP demand; Technical, network design (GPON / XGS-PON), build cost and build velocity actuals, as-builts, wayleaves and duct/pole access; Financial & tax, quality of earnings, capex capitalisation, the holdco/opco structure and capital allowances; Legal & regulatory, wayleaves and the Electronic Communications Code, street-works permits, ISP contracts, telecoms regulation and change-of-control. Lenders run their own technical and model diligence in parallel.
In real life: appoint a commercial-DD house, an independent technical adviser, a Big-Four financial/tax team, a model auditor and infra/telecom counsel, see "Who to call".
Submit the binding offer with a marked-up SPA, an equity commitment letter, committed (or highly-confident) debt and final IC approval. Negotiation centres on the price mechanism (locked-box in a competitive auction, fixing the price to a locked-box date with a no-leakage covenant, vs completion accounts), warranties and W&I insurance, the capex/build commitments, management incentives, and the conditions precedent. The seller may grant exclusivity to the preferred bidder.
In real life: W&I is placed through a broker; management is locked in with a co-invest/ratchet because the build plan is the value.
Clear the conditions precedent: merger control, telecoms-sector consent and foreign-investment screening (in the UK, the NSIA), and lender conditions. Reach financial close on the debt, the facilities agreement, security and hedging, and complete with the SPA and debt interconditional so equity and debt fund simultaneously per the funds flow.
In real life: FDI/NSIA clearance on communications infrastructure is a real timetable item, build it into the long-stop date.
Now the work that makes the money: run the build machine to plan, accelerate penetration (ISP onboarding, copper switch-off tailwinds, marketing), drive opex efficiency, execute bolt-on acquisitions of neighbouring altnets, and refinance the expensive greenfield facility into cheaper, longer debt as the EBITDA appears (next section).
In real life: the 100-day plan installs the board, the KPI pack (homes passed/connected, penetration by cohort, build cost/premise, ARPU, churn, DSCR) and the bolt-on pipeline.
Once built and penetrated, the network re-rates from a development asset onto a mature EBITDA multiple. Exit to a larger infrastructure fund or strategic consolidator, sell down to a long-hold/open-ended vehicle, recapitalise, or, at scale, IPO. Consolidation is the most common destination; the recent UK wave (e.g. CityFibre acquiring Connexin; the Jurassic / Swish / Giganet merger into AllPoints Fibre) is exactly this arc.
In real life: re-appoint a sell-side adviser ~12 months out and dress the asset, clean cohorts, contracted ISP revenue, a fundable build pipeline for the next owner.
Fibre debt evolves with the asset. You cannot raise long-dated investment-grade money against a network that does not yet generate cash, so the structure is built in stages, and getting the staging right is most of the financing skill.
At entry you raise a capex facility at the operating company, a term loan plus a committed capex line and a revolver, drawn in tranches against build milestones during an availability period. Because there is little EBITDA early, lenders size it on the business plan, not trailing earnings, and protect themselves with structure rather than coverage:
Where the senior facility cannot reach, the gap is filled with holdco debt or PIK from infrastructure debt funds, structurally subordinated to the opco lenders, and the rest with equity (the fund’s cheque plus co-invest, sometimes smoothed by an equity bridge).
| Borrower | Opco (ring-fenced), under a holdco/opco structure |
| Facilities | Term loan + committed capex facility + RCF (+ DSRA facility) |
| Tenor | 5–7 years (greenfield); refinanced longer once stabilised |
| Availability | Capex line drawn over a 3–4 year availability period against build milestones |
| Margin | Reference rate + a margin on a ratchet that steps down as leverage falls; commitment fee ~35–40% of margin on undrawn |
| Sizing | Min DSCR ~1.5–2.0× on projected mature cash flow; max gearing / LTV against mature value |
| Covenants | DSCR & gearing tests; homes-passed / penetration milestones; equity cure right |
| Cash control | Cash sweep before distributions; 6-month DSRA; mandatory interest-rate hedging (~75%+) |
| Security | Full debenture, share pledges, account & wayleave assignments; intercreditor with any holdco/PIK lenders |
As homes fill up and EBITDA appears, you term out the expensive greenfield debt into progressively cheaper, longer money, each step lowering the cost of capital and releasing equity:
| As the asset… | Refinance into… | Provided by |
|---|---|---|
| is mid-build, partly penetrated | a larger bank-and-fund capex / term facility | infra banks + debt funds, club or underwritten |
| is built & cash-generative | a Term Loan B or a long-dated bank facility | the leveraged-loan and infra-bank market |
| has stable, contracted cash flow | a US private placement or infrastructure bond | US insurers; institutional bond buyers |
| is at scale & investment-grade-like | a whole-business securitisation | a structured, ring-fenced bond platform |
You mandate a debt advisory team to run this: build the lender model and information memorandum, engage the rating agencies, and run a lender process (club or underwrite) to financial close. A refinancing as the asset de-risks is a deliberate part of the return, model it explicitly in the acquisition model’s refinancing toggle.
A deal this size is run by a syndicate of advisers and capital providers. The categories below are the ones you mandate, with illustrative examples of firms active in infrastructure and digital, examples, not endorsements or advice; pick on relationship, sector credentials and conflicts.
Runs your buy-side bid (or your build financing), the infrastructure / telecom-infra teams at the banks, or a specialist boutique.
The overbuild map, penetration-curve and ARPU view, the single most important report.
Network architecture, build cost & velocity, as-builts, wayleaves, also serves the lenders.
Quality of earnings, structuring, and an independent audit of the model everyone relies on.
SPA, financing, wayleaves/PIA, regulatory and FDI, infra/projects and telecom teams.
Places warranty & indemnity cover and arranges construction/operational insurance.
Mandate a debt advisory team to run the financing, build the lender model, engage the agencies and arrange the club. The lenders change as the asset matures:
The senior facility and capex line, the project & infrastructure finance desks.
Stretched senior, holdco and mezzanine, longer-dated, more flexible, more expensive.
Long-tenor, lower-cost capital for nationally-important coverage; export credit where vendor finance applies.
Your fund’s commingled equity, plus co-invest syndicated to LPs (pension, insurance and sovereign funds) to manage concentration, and sometimes a strategic or anchor-ISP partner.
PON equipment vendors (e.g. Nokia, Adtran, Calix), civils and build subcontractors, and an in-house or contracted build-management team, the engine of the whole plan.
The incumbent’s duct-and-pole access (Openreach PIA in the UK), the regulator (Ofcom) on the rules, and anchor ISPs contracted to drive penetration on day one.
The IC does not debate the base case, it debates the downside. The point estimate (especially for building) is the easy part; the work is sizing what goes wrong and what protects you. The sensitivities that move a fibre case most, in rough order:
| Sensitivity | Why it bites | What protects you |
|---|---|---|
| Overbuild / penetration miss | A 10pt lower steady penetration can halve the equity return, fixed cost, less revenue | Defensible (sole-fibre) footprint, anchor-ISP volume commitments, conservative ramp underwriting |
| Build-cost overrun & delay | Higher cost / home and a slower ramp deepen and lengthen the J-curve, and the negative carry | Fixed-price build contracts with KPIs/LDs, PIA to cut civils, capex contingency, milestone covenants |
| ARPU compression | Price competition from the incumbent or rival altnets erodes ARPU and the EBITDA per home | Wholesale contracts, service quality, switching costs, copper switch-off tailwind |
| Rates & refinancing | Higher rates raise the cost of the floating greenfield debt and the exit cap rate | Mandatory hedging, a conservative exit multiple, the DSRA buffer |
| Exit multiple contraction | You build at cost but exit on a multiple that may have de-rated | Underwrite a flat-to-lower exit; rely on EBITDA growth and deleveraging, not re-rating |
The IC pack therefore carries a base / management / downside case set, one- and two-way sensitivity tables on penetration × build cost and entry × exit multiple, a break-even penetration (the take-up at which the equity return hits the hurdle), and covenant headroom through the trough (how far penetration can miss before a DSCR breach forces an equity cure). Genuine downside protection in fibre comes from contracted anchor revenue, any Project Gigabit subsidy on rural premises, the copper switch-off tailwind, and the simple fact that an over-levered builder’s footprint is itself a scarce, financeable asset a consolidator will want. Run the two-way sensitivity yourself on the acquisition model.
There is no single answer, the comparator shows the trade-off moving with the assumptions. As a rule of thumb:
When there is genuine defensible whitespace, you have or can hire a proven build machine, and your LPs will fund a J-curve for a higher return. Highest value capture, highest execution risk, the return is earned.
When you need scale and cash flow now, want to avoid build risk, and can win a competitive auction at a price that still clears your hurdle. You pay for certainty, lower ceiling, lower variance.
The route most fibre platforms actually take: acquire a credible sub-scale altnet for its team, systems and footprint, then fund the organic build and roll up neighbours. Entry discipline plus build upside, if management can really execute.
In practice a platform often does all three over a fund’s life, enter via a buy-and-build, accelerate with organic build where it has whitespace, and bolt on full-network acquisitions to consolidate, then exit the combined platform to the next, larger owner. That consolidation arc is the fibre investment story of the decade, and it is playing out across the UK altnet market right now.
Market context (build costs, the CityFibre financing, the January 2027 PSTN copper switch-off, Project Gigabit, and the altnet consolidation wave) is drawn from public company, regulator and trade-press disclosure and is illustrative. Unit economics are grounded in the site’s fibre networks reference. The comparator is a simplified model, it omits churn detail, working capital, the full tax structure, refinancing and construction phasing, and is for illustration, not a forecast of any specific network. Named firms are illustrative examples of the categories of adviser and lender active in infrastructure; they are not recommendations, endorsements or investment, legal, tax or financing advice. Always take your own professional advice on a live transaction.