
Four sites, one dispatch layer. 1.9 GW nameplate across wind, tidal and tracking solar.
Siting, civil works, turbine supply, storage, interconnect, twenty-year service.
Direct-drive nacelles, immersed rotors, bifacial glass, iron-air banks, LiDAR masts.
Machines built to survive the place they stand in, and still hold the load at 3 a.m.
Aerion needed generation that could work as hard as the weather it stands in. Blades ice, screens foul, glass hazes — the machines had to survive all of that and still hand the dispatcher the same number every morning.
Built the portfolio as interlocking parts: ridge wind for the shoulder hours, the tide race for the base, tracking solar for the peak, iron-air banks to carry the gap. Each site is sized against the others, never against a brochure.
Surveyed on foot before anything was drawn. Half Moon Shoal, the tide race, the open meadow. No borrowed wind atlas, no stand-in soil report. The data came from inside the landscape the machines were built for.
The portfolio was sized against published numbers, not a pitch deck. Three of them set the frame this case file was drawn inside.
Renewables led by solar, plus nuclear, are forecast to reach 50% of global electricity generation through 2030, up from 42% in 2025, while coal falls to 27% from 34%. Everything here is built for the half of the grid that has to hold when the fossil half is gone.
Renewables are projected to move from 32% of global generation in 2024 to 43% by 2030, with the variable share almost doubling to 27% and curtailment already rising in China, Germany, Brazil, Chile, the UK and Ireland. Storage and dispatch were priced in from the first drawing for that reason.
In 2024 renewables supplied about a third of world electricity: hydropower 14%, wind 8%, solar PV 7%. A portfolio that leans on one of the three is a portfolio with a season it cannot cover.
Salt takes coatings. Grit takes bearings. The machines had to survive both and still read on the meter at three in the morning.
The portfolio holds ridge wind, tidal race, tracking solar and iron-air storage. Range comes before spectacle. Each asset earns its place against the hours the others cannot cover.



Every site surveyed on location. Ridge line, tide race, open meadow. The machines needed ground that didn't fight them.
No borrowed atlas. No stand-in soil report. Same hands on the survey, same hands on the commissioning sheet.


Heritage ground, modern load

Coatings, alloys, glass and cable. Every colour, every face, every spec tested against the weather before it earned a place in the system.
Neiko carries the wordmark and every headline. Hugos works only in tracked caps, at label and signage size, so the two never compete. An old-style serif holds the editorial voice, and the mark constructs on a geometric grid. Every choice traces back to where the power lands.
Flat-sided display face, drawn more than typeset. Carries the wordmark and every headline. Holds at stamp size and at full bleed alike.
Only ever set in wide caps: nacelle plates, gate signage, section markers. Never runs as a headline, so it never competes with Neiko.
Light geometric sans. Runs long without shouting. Abstracts, offtake terms, service manuals.
Old-style serif for the lead paragraph, pull quotes and certifications. The least precious face in the system.
System monospace for capacities, coordinates and tags. Small, uppercase, tracked wide.

Survey, civil, supply, commissioning, service. Same hands from the first met mast through the last dispatch report.
Aerion asked for generation that could survive where the power gets used. The system holds at every hour, and the survey came from the same ground the machines stand on.
Every figure in this file traces to a published source. Where our measured number differs from the literature, the measured number is the one printed.
Ridge wind, coastal wind, and the towers that carry it. The assets in this section are onshore ridge lines and near-shore platforms. The work covers met-mast campaigns, hub-height selection, foundation design, and the commissioning record a grid operator signs off from.





Hub-height wind is extrapolated up or down from a 110 m reference using a power-law shear exponent of 0.2, and annual energy production comes from a Weibull distribution fitted per wind class. Our met campaign follows the same method so the numbers can be argued with.
Turbine specific power is expected to keep falling toward roughly 150 W/m², trading rotor area against rating. Rotor diameter, specific power and hub height can each be traded to reach a given capacity factor.
Upsizing machine and plant together moves cost of energy more than either alone, which is why the layout was priced at three hub heights before the foundations were drawn.
Two LiDAR masts, eighteen months, before any hub height was fixed.
Row spacing modelled against the prevailing sector, not the property line.
Gravity bases on rock, piled where the till runs deep.
Power curve verified machine by machine against the contract curve.
Wind power goes as the cube of speed. Ten per cent more wind is a third more power, which is why the mast comes before the drawing.
Tidal races, immersed rotors, and the base hours nothing else covers. The assets in this section run on water that moves whether the wind blows or not. The work covers bathymetric survey, screen and rotor design, marine civils, and forty years of maintenance planning.





Four 1.5 MW turbines, 6 MW installed, on gravity-base foundations in the Inner Sound; 10.2 GWh net in 2023, and the first tidal stream array anywhere to pass 50 GWh. The machine sheet above follows this reference class.
The national practical resource is put at about 34 TWh a year, near 11% of UK electricity demand, while operating projects have so far shown inclusive capacity factors below 0.2. We size against the demonstrated figure, not the theoretical one.
Rated 1.5 MW at 3.0 m/s with an 18 m rotor, upstream and fully submerged. Every rotor in our race is specified to be recoverable afloat, because the ones that cannot be serviced stop earning.
Multibeam survey of the channel floor before any rotor position was picked.
Twelve months of ADCP data, calibrated against the published tide table.
Ballasted bases placed in the slack hour, cable trenched to the shore landing.
Rotors lifted afloat on a forty-year cycle, screens cleared each spring.
The tide is the only fuel with a published timetable. It is not always there, but it is never a surprise.
Tracking arrays, bifacial glass, and the peak they were built for. The assets in this section are single-axis tracking fields on grazing ground and old meadow. The work covers yield modelling, row pitch, glass selection, and the storage that carries the array past sundown.




Bifacial modules on single-axis trackers over natural ground cover, albedo 0.2 to 0.3, generally see gains under 10%. Our 6.8% measured gain sits where the literature says it should, and the model was built to that ceiling.
Measured against a monofacial reference at 10 m axis spacing, albedos of 19%, 32% and 63% returned energy gains of 7.9%, 11.9% and 19.2%. Ground treatment, not glass, is the lever.
Raising ground albedo to 0.5 lifts gain toward 20% globally, and elevating modules a metre off the ground recovers self-shading losses toward 30%. Both were priced here; the flock won.
The iron-air chemistry is built for about 100 hours of discharge at rated power — against roughly four for lithium-ion — at a fraction of its cost per kWh, with 10 MW / 1,000 MWh systems in build. Ours is sized at four hours because that is what this grid pays for today.
Eight configurations run against measured albedo, not a default value.
Pitch set at 4.2 m to keep the sheep, the mower and the shading honest.
Hail rating and bifaciality floor written into the supply contract.
Four hours of iron-air behind the fence, charged off the midday clip. The chemistry runs to a hundred.
The array is sized for the afternoon, and the battery is sized for the argument that follows it.