Reliable, affordable, sovereign and clean. A proposed direction for Isle of Man electricity: solar, storage, household flexibility, two interconnectors. High optionality, no single bet, tested against the island's asset clock, debt and planning politics. A proposition to argue with, not a plan to obey.
The decision is forced either way. The island's two keystone assets retire within a decade of each other: the Pulrose gas plant, roughly 90% of our electricity, around 2033; the single subsea cable to England, in service since 2000, around 2035. Doing nothing is not on the table. The question is what replaces them.
The proposal is a portfolio, not a monument. Two interconnectors of around 100 MW, diversified in route and ideally counterparty. A grid-forming utility battery of 30 to 40 MW for frequency support, fast reserve and black-start. 45 to 60 MW of solar across rooftops, businesses and a few well-sited ground arrays, with community batteries and, later, bidirectional EV charging. Efficiency first, because the cheapest unit is the one never bought. Gas stays as standby until the new kit has proven itself, then retires. Every element is independently useful. High optionality, no single bet.
On the planning record, this is the buildable path. Rooftops, substations, industrial land, seabed. None of it needs the turbines the island has organised against for three years. Wind stays an option; nothing here depends on winning that fight.
And it is financially grounded. Manx Utilities carries a £185m bond to 2034 on 85,000 people paying 29.5p a unit. The proposal phases £320–420m over 15 years, mostly bankable infrastructure, and avoids the El Hierro trap of chasing total self-sufficiency at any cost. Jersey lands power at 22.20p by importing most of it. Sovereignty comes from redundancy, not from generating every electron at home.
One caveat before anything else. This was put together by an engineer and an AI, not a power systems consultancy. Treat it as a costed direction with the homework shown, and treat commissioning real grid engineering as the first action inside it.
Everything below is the measured starting point, not a scenario.
The Pulrose CCGT (25–85 MW) supplies about 90% of island electricity, end of life ~2033. The sole interconnector, a 104 km AC cable rated 40–60 MW, follows ~2035 on the independent system guide's estimate; Manx Utilities suggests it could run longer, but a 1990s cable carrying the island's future deserves a funded decision, not optimism. Behind them: a ~38 MW diesel station at Peel, ~50 MW of reserve diesels at Douglas, 1.2 MW of hydro and the 5.5 MW Energy from Waste plant. Storm Éowyn cut power across the island in January 2025 and damaged Peel station on its way through.
Policy already points this way: Tynwald adopted 100% decarbonised electricity by 2030 back in 2022, and Manx Utilities' roadmap pairs local renewables with a new interconnector. Delivery has not followed: the 2030 pledge is officially "at risk", the 30 MW renewables target has produced ~7 MW of private rooftops, the wind farm is still pre-application, and the second interconnector is a £200m estimate with earliest commissioning ~2031. The proposal below is about closing that gap with tools the island will actually consent to.
Build your own mix and see what it does to energy balance, winter security, carbon and indicative cost. The model is deliberately simple and every assumption is listed below it; treat outputs as directional, not as a system study.
Eight questions any serious energy plan must answer, taken in turn.
Every element runs today on a grid this size. Kauai (~82 MW peak against our 75) operates at up to 90% inverter-based generation and rode through the trip of a 26 MW unit in hundredths of a second. Suðuroy ran ten days on 100% inverter-based wind with a 6.3 MW battery and a synchronous condenser. Shetland hangs an archipelago off one HVDC link with a 68 MW bridging battery. Nothing here makes the island a guinea pig.
The caveat is solar physics. At this latitude solar manages a 10–11% capacity factor and near-nothing on a January evening; 50 MW yields roughly an eighth of annual demand, mostly in summer when demand bottoms at 23 MW. Solar plus storage cannot run a Manx winter, which is why the interconnectors and gas standby are load-bearing parts of the design, not accessories.
Today the grid's stability is Pulrose: spinning steel providing inertia, fault current and a frequency reference. Retire it without a plan and protection relays stop tripping, because inverters deliver only 1.1–1.5 times rated current into a fault. The field-proven 2030s architecture: two or three separated grid-forming battery stations (30–40 MW) as the fast backbone, one or two flywheel-boosted synchronous condensers for fault level and voltage, and protection reset to a higher rate-of-change-of-frequency standard, as Ireland did. Exactly the territory where this proposal needs real power-systems engineering, not an explorer widget.
Black-start precedents exist: Hawaii's Kapolei battery is contracted to restart thermal plant after a blackout, and Kauai's battery powered an intentional island including hospital loads in 2024. Grid-forming BESS plus retained gas as the diverse second path beats today's position, where restoration hangs on one thermal site and one cable. The unforgiving rule: fast reserve must always cover the largest single infeed. Two links plus batteries keep that number manageable; one big link without storage does not.
On 2025–26 UK benchmarks: ground solar ~£0.65–0.8m per MW, utility BESS ~£650k per MW plus 3–8% for grid-forming, and Manx Utilities' own £200m estimate per interconnector. The whole portfolio lands indicatively at £320–420m over 15 years. Reference points the island already knows: £40–46.5m for one 20 MW wind farm, £30m for the consented Billown solar farm, £660m for Shetland's single cable.
Operating costs run the other way and pay for the capital: every imported or solar kWh displaces gas bought on the market that handed Manx Utilities a £42m loss in 2022–23 and customers a near-60% price shock. Jersey, importing ~95%, charges 22.20p against our 29.5p. Interconnector systems trade fuel risk for financing risk, and financing risk can be fixed for 25 years at signing.
Cable fault: Jersey's EDF1 failed permanently in 2012 (weeks of diesel, a 9.5% tariff rise); Guernsey's single link failed in 2018 and hurt its finances badly. Two Manx links on separate routes leave 100 MW plus batteries plus standby; today the same fault leaves one gas plant that retires in 2033. Fuel disruption: the island's gas arrives through one 11.7 km subsea spur; every megawatt of solar, storage and interconnection shrinks that exposure. Peak demand: electrification pushes peak toward ~130 MW by 2050; UK trials show 8–15% peak reduction from flexibility, and V2G turns the car fleet into distributed capacity.
A bonus the island's network engineers will recognise: subsea power cables can bundle fibre, as today's link already does (operated by e-llan within Manx Utilities). Fibre pairs in both new links buy resilient connectivity alongside resilient power.
Counterparty diversity is not hypothetical: France threatened to cut Jersey's power in the 2021 fishing dispute. One link to Britain and one to Ireland is the strongest sovereignty position available; two separated British routes still solve the engineering. The Irish option earns a feasibility study, not an assumption. The bolder configuration, a GB–Mann–Ireland route, could in principle sell spare transit capacity between the two power markets, the way commercial interconnectors earn their keep; that belongs in the study's scope.
Manx households pay 29.5p; import-led Jersey pays 22.20p. The gap is mostly fuel exposure, debt service and subscale thermal generation. The realistic ambition is a decade of below-inflation rises, then genuine cuts as the bond rolls off in 2034 and fuel purchases shrink.
On standing charges: the island already runs a low 25.5p/day charge that Manx Utilities admits under-recovers connection costs. A mostly-fixed model would blunt every incentive this proposal depends on. The coherent design is the opposite: low honest fixed charge, meaningful unit rate, visible credits for flexibility. An export rate pegged to avoided import cost (Jersey recalculates its 9.53p buy-back annually on exactly that basis), cheap EV windows, paid peak events, a social tariff. The smart meter fleet, due complete end-2026, makes all of it cheap to run.
One government-owned utility with Tynwald-approved pricing can capture the whole value of flexibility without Britain's supplier-versus-network coordination mess. What is missing is the rulebook: a formula-based export tariff instead of a unilateral one; time-of-use as the smart-meter default with an EV/V2G window; a standing flexibility register with posted prices; connection standards for home batteries and bidirectional chargers; and independent economic oversight of electricity, most simply by extending CURA's existing gas remit. Most of this is tariff policy Tynwald already approves annually.
The cheapest megawatt is the one never generated. The Energy Efficiency Scheme has reached ~4,500 households; full uptake saves ~£700 a year per home. UK evidence puts 8–15% peak reduction in reach from time-of-use and smart heating. Heat Wise and the new-build boiler ban will electrify heating and raise winter peak, so every insulated loft shrinks the cables, batteries and standby the island must finance. Efficiency is infrastructure. Fund it first.
Set out in full in the phased plan below: prove and prepare to 2031, replace and switch to 2036, optimise and retire thereafter. The sequencing rule throughout is that no legacy asset is withdrawn until its replacement has operated through a winter.
Opposition to wind on the island is a fact of the planning record, not a prejudice to be lectured away. This strategy treats it as a design constraint.
The record is specific. The publicly-owned Earystane/Scard wind farm ("Cair Vie", turbines to 185 m, £40–46.5m) was meant to be running by end-2026; in July 2026 it had not yet filed a planning application, against an organised community campaign, a defeated Tynwald motion and a Tynwald Day petition. Offshore, Ørsted's 1.4 GW Mooir Vannin faces a 5,000-signature opposition petition and an MoD radar objection, with a Council of Ministers decision expected 2027. Yet consultations show support for renewables in principle: 80% backed onshore wind in 2019. The pattern: support in the abstract, contention at every specific site.
Compare the alternatives. Rooftop solar sits on buildings; ~1,000 households already have it. Batteries live in containers on industrial land. Interconnectors are invisible once laid. Flexibility is a tariff. Efficiency is money in pockets. Only ground solar touches the landscape debate at all, and Billown drew nothing like wind's resistance. None of this says wind is bad; Manx wind above 10.5 m/s is a superb resource. It says energy security should not queue behind the island's most contested planning fight, and a plan that works without wind can always add it later.
Mooir Vannin, honestly: an export project in Manx waters, its value to the island fiscal rather than electrical, its fate with the examination panel and Council of Ministers in 2027. Under Attorney General guidance, election candidates are constrained from campaigning on it. This page treats it as a scenario: if consented, negotiate hard on option fees and a future island connection; if refused, nothing here changes.
| Option | Visible footprint | Organised opposition to date | Consent path | Political risk |
|---|---|---|---|---|
| Rooftop / commercial solar | None beyond the building | None recorded | Largely permitted development | Low |
| Community & utility batteries | Containers on industrial land | None recorded | Standard planning | Low |
| Interconnectors | Landfall works, then none | None recorded; broad support in MU consultation | Marine consent + landfall planning | Low–medium |
| Flexible demand / V2G / efficiency | None | None; standing-charge politics only | Tariff policy via Tynwald | Low |
| Ground-mount solar | Fenced lowland arrays | Minimal (Billown) | Standard planning; consented precedent | Low–medium |
| Onshore wind | Turbines to 185 m on uplands | Organised (Earystane/Scard CAG); Tynwald motion; petitions | Contested; still pre-application after 3+ years | High |
| Offshore wind (Mooir Vannin) | 87 turbines, 6–12 nm offshore | Organised (NOT4IOM, 5,000+ signatures); MoD objection | MIMA examination; CoMin decision 2027 | High, and outside island control |
The main building blocks compared honestly, including the ones this strategy does not lead with.
| Option | For | Against | Role here |
|---|---|---|---|
| Rooftop & commercial solar | Zero landscape cost; builds public ownership of the transition; displaces daytime imports; ~1,000 households already in | 10–11% capacity factor; near-zero in winter; island install costs above UK mainland; needs storage or export headroom to be useful at scale | Core, 25–30 MW by 2035 |
| Ground-mount solar | Cheapest solar per kW (~£0.65–0.8m/MW); Billown proves consentability; fast to build | Land take on a small island; same winter problem; grid connection depth at 33 kV (~25 MW sections) | Core, 20–30 MW across 2–3 sites |
| Community batteries | Soaks up local rooftop surplus; defers network reinforcement; visible neighbourhood benefit | ~£290/kWh installed; small individually; needs a tariff structure that rewards them (does not exist yet) | Supporting, 10–20 MWh |
| Utility grid-forming BESS | Frequency response, virtual inertia, black-start, islanding, all field-proven at this grid size; ~£650k/MW plus 3–8% grid-forming premium; the keystone stability asset once Pulrose retires | Two-hour duration is power security, not winter energy; revenues need a designed island market, not GB merchant assumptions; fire-safety siting diligence | Core, 30–40 MW / 60–80 MWh in 2–3 stations |
| V2G / bidirectional charging | Turns the car fleet into distributed storage; UK trials show £360–620/yr value per vehicle; AC hardware costs falling toward £1–2k | Only ~1,377 EVs on-island (2024); vehicle/charger compatibility still narrow; a late-2020s scaling play, not a near-term firm resource | Supporting, scale from ~2029 |
| Two interconnectors | Decarbonised imports at Jersey-grade cost; N-1 security against cable fault; export route for surplus solar; can bundle dark fibre as today's cable does; ~£200m each on MU's estimate | Biggest single capital items; delivery risk (cable ships are booked years out); import dependence needs managing via contracts and diversity | Core, replace by ~2032 + second link by ~2035 |
| Transitional gas / diesel standby | Cheapest firm winter capacity the island already owns; fuel-diverse black-start second path; Jersey and Shetland both keep standby thermal behind their links | Fuel price volatility caused the 2022–23 shock; carbon; Pulrose itself expires ~2033 so standby needs its own plan | Transitional, shrinking to ~40 MW standby, retire when proven redundant |
| Onshore wind | Superb resource (>10.5 m/s); 20 MW would cut winter imports meaningfully; complements solar seasonally | The one option with organised, sustained opposition; still pre-application after 3+ years; £40–46.5m for 20 MW; strategy must not depend on it | Option, not a dependency |
| Offshore wind (Mooir Vannin) | £2bn claimed economic value over 35 years; no island capital at risk; potential future connection option | An export project: island keeps rent, not power; contested; decision (2027) outside island control and constrained for election candidates | Scenario; negotiate value hard if consented |
| Efficiency & demand reduction | Cheapest option per MW; £700/yr savings per fully-treated home; shrinks every other asset; UK trials show 8–15% peak reduction | Retrofit delivery capacity on-island is thin; benefits diffuse and slow to aggregate; needs sustained funding, not pilots | Core, first money spent |
Sized for the 2035 system: ~95 MW winter peak after efficiency, ~430 GWh, Pulrose retired, both links live.
| Asset | Size | Indicative capital | Phase | What it buys |
|---|---|---|---|---|
| Efficiency & flexibility programme | 8–12% peak reduction | £25–40m | 2026–2031 | Shrinks every asset below; bills fall first |
| Interconnector 1 (UK, replacement) | 100 MW | ~£200m | order by 2028, live ~2032 | Replaces the 2000 cable before end of life |
| Utility grid-forming BESS | 30–40 MW / 60–80 MWh, 2–3 sites | £25–35m | first station by 2029 | Frequency, fast reserve, black-start, islanding |
| Synchronous condensers | 2 × 10–20 MVA, flywheel-boosted | £15–25m | with Pulrose retirement | Fault level and protection security |
| Rooftop + commercial solar | 25–30 MW | £25–35m (mostly private capital) | continuous | Citizen-owned daytime energy |
| Ground-mount solar | 20–30 MW, 2–3 sites | £15–25m (private/PPA) | 2027–2032 | Cheapest island generation |
| Community batteries | 10–20 MWh | £3–6m | from 2028 | Local absorption, network deferral |
| V2G enablement | 3,000–5,000 vehicles by 2036 | £5–10m (chargers, mostly private) | scale from 2029 | Distributed peak capacity |
| Interconnector 2 (Ireland preferred, UK fallback) | 100 MW | ~£200m | FID by 2030, live ~2035 | N-1 security; counterparty diversity; export headroom; second fibre route |
| Gas/diesel standby | ~40 MW retained | £10–20m (life-extension) | until both links + BESS proven | Resilience of last resort; retire ~2040 |
Total programme: indicatively £320–420m of which roughly £120–160m is private or householder capital. For scale: Shetland's single new link alone cost £660m, and Manx Utilities' 2017 debt restructuring wrote off £95m. Financing via long-dated utility borrowing against regulated revenues, UK/Ireland interconnector frameworks where applicable, and PPAs for solar. No unlimited public funding is assumed anywhere in this plan.
| Risk | Likelihood / impact | Mitigation |
|---|---|---|
| Interconnector delay (cable ships, consents, converter lead times) | High / High | Order Link 1 by 2028; keep Pulrose maintainable to 2033+; stage BESS early so the island can ride N-1 during the gap |
| Cable fault once import-led | Medium / High | Two links on separated routes and landfalls; BESS sized to largest infeed; 40 MW standby thermal held through the 2030s; spares and repair contracts pre-agreed (Jersey's playbook) |
| Import price exposure replaces gas price exposure | Medium / Medium | Long-term contracts on the Jersey/EDF model; fixed-price solar PPAs on-island (Kauai's hedge); retain export revenue from summer surplus |
| Financing strain on a small utility already carrying £185m | Medium / High | Phase capital; recycle the bond roll-off from 2034; private capital for solar and chargers; do not gold-plate (one BESS fleet, not one of everything) |
| Stability shortfall as Pulrose retires (inertia, fault level, protection) | Medium / High | Grid-forming spec mandatory in BESS procurement; synchronous condensers commissioned before retirement; RoCoF protection programme; islanding trials each year |
| Public backlash if bills rise to fund the build | Medium / High | Efficiency first so bills fall before capital lands; transparent formula-based tariffs; social tariff; publish the avoided-fuel arithmetic annually |
| V2G / flexibility underdelivers | Medium / Low | Treat V2G as upside, not baseline; the N-1 math above never counts more than 30% of enrolled vehicles |
| Solar consenting friction at ground-mount sites | Low / Medium | Site selection on previously-developed and low-grade land; community benefit sharing; Billown precedent |
| Gas spur or fuel logistics failure during transition | Low / High | Dual-fuel standby capability; strategic fuel storage at Peel maintained until Link 2 is live |
A strategy is only as good as its first phone call. Here is what has to happen now, and the futures the plan must be able to absorb.
The most time-critical action costs almost nothing: open formal connection discussions with National Grid. A 100 MW link needs a GB-side connection agreement, and connections, converters and cable ships are queued years deep. "Earliest ~2031" is only true if the conversation starts now; every year of delay stretches the exposure window when Pulrose is gone and one 35-year-old cable is doing the work. The parallel call to EirGrid buys the island's strongest sovereignty option for the price of a feasibility study. Alongside: commission an independent power-systems study to kick the tyres on everything on this page, issue the grid-forming BESS specification, publish the flexibility rulebook, and get consented ground solar moving. None of this waits for an election.
A hyperscale AI campus runs near-flat: 30 MW adds ~40% to average demand, 100 MW more than doubles it. Try the slider above. The opportunity is real: an anchor tenant could underwrite the second interconnector the way Viking underwrote Shetland's £660m link, and a datacentre plus two fibre-equipped cables is an economic strategy, not just an energy one. The sequencing is unforgiving: operators shop on time-to-power, and an island that has not opened its National Grid conversation has nothing to sell. Decide in advance what to ask of such a tenant: heat reuse, flexibility commitments, a contribution to firm capacity.
If Mooir Vannin is consented, the island's negotiating position changes and nothing else. If gas prices spike early, efficiency, the first battery and the existing cable are the cushion, which is why they come first. If the old cable fails before replacement, the island runs on Pulrose and diesels while a repair mobilises: survivable today, catastrophic after 2033, and the strongest argument for ordering Link 1 by 2028. If a future Tynwald wins genuine consent for wind, 20 MW slots straight in and the plan improves.
Distributed flexibility plus strong interconnectors looks stronger than a generation-heavy model, and not by a little. Generation-heavy self-sufficiency at 85,000 people means either the wind fight the island has spent three years not winning, or El Hierro economics: bespoke kit at €6,000–7,000 per kW that peaked at 54% of annual supply. Pure import dependence fails the other way, as Jersey's 2012 cable failure and the 2021 French threat showed. The portfolio sits between: interconnection for cost and carbon, distributed solar and storage for sovereignty, flexibility because it is the cheapest capacity, standby thermal until redundancy is proven. Kauai runs this model at our scale. Jersey prices it. Shetland financed it. But this is a proposition, not a verdict: it needs the power-systems engineering the Observatory does not have, the explorer exists so you can build the counter-case, and the best counter-cases get published.
Everything the explorer computes, stated plainly so it can be argued with.
Hidden in Quick read. Switch to Full brief for the assumption cards.
2026 baseline 360 GWh / 75 MW winter peak / 23 MW summer minimum (Manx system guide). Horizons: 2030 ≈385 GWh, 82 MW; 2035 ≈430 GWh, 95 MW; 2045 ≈500 GWh, 115 MW, an Observatory interpolation toward the ~130 MW by 2050 projection as heat and transport electrify. Monthly demand shape winter-weighted. Flexibility slider reduces peak only, plus a small energy saving. The large-load slider adds a near-flat demand block (90% utilisation, 95% coincident with peak) that flexibility does not touch.
Solar: 10.5% annual capacity factor, UK monthly profile, so 1 MW ≈ 0.92 GWh/yr, heavily summer-weighted. Hydro 4 GWh. Energy from Waste (the incinerator) 5.5 MW / ~32 GWh, assumed continuing. Existing cable 40 MW firm (60 MW sustained) to 2035. New links 100 MW each. BESS shifts energy within days; it is counted as firm capacity, not annual energy. V2G firm contribution = vehicles × 7 kW × 30%. Gas standby is firm capacity and the residual energy supplier when imports and solar fall short.
Capex: ground solar £0.75m/MW; rooftop blend £1.2m/MW; utility BESS £0.68m/MW (2h, grid-forming); community storage £290/kWh; V2G charger £1,500; interconnector £200m per 100 MW link (MU estimate); syncons £20m; efficiency £3,000 per treated home. Indicative unit-cost delta: annuitised capex (6%, 25y cables/solar, 15y batteries) plus fuel/import costs (imports 9p/kWh delivered, gas 8p/kWh fuel, solar 0 marginal) divided by units sold, compared against a counterfactual that keeps the gas model, which after 2033–35 still means replacing the CCGT (~£80m) and the cable (~£200m). Fuel price volatility, the thing that produced the 2023 shock, is not priced in on either side. Directional only.
Typical design lives: subsea cables ~40 years (the existing link will have done ~35), HVDC converter electronics ~20 before refresh, solar panels 25–35 with inverter swaps every 10–15, grid batteries 15–20 with mid-life augmentation as capacity fades, synchronous condensers 30–40, V2G chargers 10–15, insulation 40+, CCGTs ~30 (hence Pulrose's 2033). The model annuitises in two buckets: 25 years at 6% (cables, solar, syncons, efficiency) and 15 years (batteries, chargers, gas life-extension). That treats cables conservatively (25-year money on a 40-year asset) and batteries slightly generously (no augmentation line); the errors pull opposite ways, and a real financing case would replace both with actual tenors.
Grid intensity: gas generation 400 g/kWh; imports 100 g (2030), 50 g (2035), 25 g (2045) reflecting GB/Irish decarbonisation trajectories; solar and hydro 0 (lifecycle emissions excluded consistently); Energy from Waste treated as carbon-neutral here since it exists for waste disposal regardless, a simplification worth arguing about. N-1 test: firm capacity minus the largest single asset must exceed peak after flexibility. Largest asset is whichever is biggest of: one link, the BESS fleet's largest station (assumed half the fleet), or gas standby.
Primary and secondary sources behind the figures. Where sources conflict (the existing cable is quoted at 40 MW continuous and up to 60 MW sustained), the text says so.
Hidden in Quick read. Switch to Full brief for the full source list.