Eolus Aktiebolag (publ) - Deep Dive Research Report

Ticker: EOLU-B.ST (Nasdaq Stockholm, Class B) | Sector: Utilities / Renewable energy development Report date: 19 September 2026 Most recent reported period: Q2 2026 (1 April to 30 June 2026), published 27 August 2026. Next report: Q3 2026 on 13 November 2026 (Financial Calendar)

A note on sources before we begin

Eolus reports on a calendar fiscal year with four quarterly reports. Working from the published cadence (Q1 on 6 May 2026, Q2 on 27 August 2026, Q3 scheduled 13 November 2026, year-end on 16 February 2027), the most recent released period as of today is Q2 2026, published 27 August 2026 and confirmed in the company's own release. Q3 2026 is not yet due.

The six most recent reporting events used throughout this report are:

One important limitation, stated plainly: Eolus does not publish verbatim transcripts of its quarterly webcasts, and the third-party transcript archives that hold them (MarketScreener, Quartr) are subscriber-gated. What is public for each of the six events is the interim report itself including a signed CEO statement, the slide presentation, and the live webcast in Swedish with a written Q&A channel. Every management statement quoted in Sections 7, 8 and 9 below is therefore sourced to the dated interim report or the dated regulatory release that accompanied that quarter's webcast, not to a paraphrase of audio I could not obtain. Where a commitment was made verbally and is only knowable from a transcript, I say so rather than invent it.

1. What the company does

Eolus builds wind farms, solar farms and grid-scale batteries, and then sells them. That is the business in one sentence, and the selling is not incidental to it. Eolus is not a utility that owns power plants and collects electricity revenue for thirty years. It is a project developer, which means its actual product is a fully permitted, grid-connected, contracted, buildable energy asset, and its customer is an investor who wants to own that asset once the hard part is done.

The hard part is the entire business. Turning a piece of forest in Västerbotten or a patch of desert in La Paz County, Arizona into something an insurance company will wire hundreds of millions of euros for takes between five and fifteen years of unglamorous, failure-prone work: securing land leases from dozens of individual owners, running wind measurement campaigns, applying for a grid connection and waiting in a queue for it, producing environmental impact assessments, surviving municipal and national permitting, negotiating a turbine supply contract, arranging a power purchase agreement with a creditworthy offtaker, and structuring project debt. Most projects that enter that funnel never leave it. The ones that do are worth a great deal more than the paperwork they consist of, because they let a capital allocator deploy money into an operating asset without owning the risk of the decade that preceded it.

Eolus then does something that meaningfully changes the economics: it keeps the asset on its own service books after selling it. When it sold three Swedish wind parks to Mirova in December 2025, it signed a fifteen-year asset management agreement for those same parks and agreed to jointly develop battery storage at all three sites (Eolus release, 23 December 2025). So one development success produces a large lumpy cash event today and a small recurring fee stream for fifteen years, plus a fresh development site for the next technology.

The company puts its own description of the model this way on its investor pages:

"a pure-play developer that uses efficient capital recycling to create value above the cost of capital"

• Eolus investor relations

The founding story, and why it still matters

Eolus was founded in 1990 in Osby, in the far south of Sweden, by Bengt Simmingsköld, and the company describes itself as the first commercial wind power company in Sweden (History & Name). Its first installation was in 1991: two Vestas 225 kW turbines at Grönhögen on southern Öland. For scale, a single modern turbine at the Fågelås park Eolus completed in 2025 stands 250 metres to blade tip and produces many multiples of what those two 1991 machines did between them.

The company reached its hundredth turbine in 2004, the same year it moved its head office from Osby to Hässleholm, where it remains. It had built more than 600 turbines by 2020 and more than 800 by 2025 (History & Name). By the company's own account it has built roughly 13% of the wind turbines in Sweden (Markets).

Two things in that history explain the company you see today.

First, the geographic sequence. Estonia in 2006, Latvia in 2011, Norway in 2012, Finland in 2014, the United States in 2015, Poland in 2021. Each entry was a response to the home market becoming harder, and each one took years to produce a first sale. The US entry in particular is now the part of the business that generates the largest single cash events, and it took roughly seven years from entry to the first battery project sale.

Second, the technology sequence. Eolus was a wind company for its first twenty-five years, which is why it was called Eolus Vind (vind is Swedish for wind). Solar and battery storage were added in the second half of the 2010s. By June 2026 the portfolio was 5,834 MW of onshore wind, 1,000 MW of offshore wind, 4,988 MW of solar and 2,975 MW of energy storage (company site, June 2026 figures). Wind is now less than half of it. The company acknowledged this formally: the AGM on 15 May 2025 resolved to drop "Vind" from the legal name, and Eolus Vind Aktiebolag became Eolus Aktiebolag on 9 June 2025 (Eolus Vind changes name to Eolus).

The name itself comes from Aiolos, the keeper of the winds in the Odyssey, in an older Swedish spelling.

What the value proposition actually is, to whom

There are three distinct groups paying Eolus, and it is worth separating them because they buy completely different things.

Institutional asset owners buy de-risked projects. A pension fund, an insurer or an infrastructure fund has a mandate to own long-duration inflation-linked cash flows, and essentially no capability to spend eight years negotiating land leases in northern Sweden. Eolus converts a decade of permitting risk into a single closing date. Mirova, MEAG (the asset manager of Munich Re), Aquila Capital, KGAL, ewz (the municipal electricity utility of Zurich), Latvenergo and DESRI have all been on the other side of an Eolus transaction.

Electricity buyers buy volume and price certainty. Alcoa contracted the full output of the Øyfjellet wind farm in Norway for fifteen years, around 1.2 TWh a year, to power aluminium smelting (Eolus release). Google contracted the output of four small Swedish parks for ten years to supply its Finnish data centre at Hamina, announced 22 January 2014. Sonoma Clean Power, a Californian community choice aggregator, took a ten-year tolling agreement on the Pome battery in Poway, California.

Asset owners who have already bought buy operational competence. Eolus had 1,271 MW under asset management at the end of Q2 2026, up from 967 MW at the end of 2024 (Q2 2026 report; Year-End Report 2025).

A concrete walk-through: Centennial Flats

The cleanest illustration of how Eolus actually earns money is Centennial Flats, a 500 MW solar plus 267 MW battery project in La Paz County, Arizona.

Eolus acquired the project in 2018 and continued developing it: interconnection, land, permitting, offtake structuring. In 2022 it sold the project to Copia Power, a Carlyle Group portfolio company, before construction (project page). Eolus did not receive the money in 2022. The consideration was structured as an upfront payment plus a long tail of milestone payments tied to construction and commercial operation of the project's three phases, with Eolus providing support services to Copia through construction.

The tail is still running in 2026. On 14 August 2026 Eolus announced it had received approximately USD 25 million after two of the three phases reached commercial operation, and raised its forecast for remaining gross proceeds to approximately USD 12 million on third-phase commercial operation, expected before the end of 2026. Total expected revenues from the project are approximately USD 116.9 million, of which USD 110.0 million had been collected. The net effect on operating profit and cash flow was guided to approximately USD 20 million in Q3 2026 (Eolus release, 14 August 2026).

Read that sequence again, because it is the whole company. A project acquired in 2018, sold in 2022, still paying cash in 2026, on an aggregate that took eight years to collect. This is why any single quarter of Eolus tells you almost nothing. In Q2 2026 no transaction closed and no project crossed a revenue-recognition threshold, so sales were close to nothing and the quarter was a loss. Three weeks after the quarter closed, USD 25 million landed. The business is genuinely this lumpy, and the lumpiness is structural rather than a sign of anything going wrong.

2. Business segments

Eolus reports two segments: Project Development and Asset Management. They are radically different businesses that happen to be very good for each other, and the second exists largely because of the first.

2.1 Project Development

What it does. This segment originates, permits, builds and sells renewable generation and storage assets, and provides technical advisory work to other renewable energy participants. It spans onshore wind, offshore wind, utility-scale solar, standalone battery storage and hybrid (co-located) projects, across Sweden, Finland, Norway, the Baltics, Poland and the United States.

The portfolio at 30 June 2026 stood at approximately 14,797 MW (company site), split:

• Onshore wind: 5,834 MW
• Solar: 4,988 MW
• Energy storage: 2,975 MW
• Offshore wind: 1,000 MW

That last line is worth flagging immediately, because it is about to change. At 31 December 2025 the offshore line was also 1,000 MW, and the portfolio total was 15,820 MW with 6,409 MW onshore wind, 5,309 MW solar and 3,102 MW storage (Year-End Report 2025). On 16 July 2026 the Swedish government rejected the permit applications for both of Eolus's Swedish offshore projects, Västvind (roughly 1,000 MW off Gothenburg) and Najaderna (roughly 1,000 MW off Gävle and Tierp), on national defence grounds (Eolus release, 16 July 2026). Capitalised costs for Västvind of approximately SEK 48 million sat on the balance sheet at 30 June 2026 and were written down.

The core capability. The thing Eolus knows how to do that a financial buyer cannot is navigate local permitting and grid access in jurisdictions where the deciding authority is often a municipal council with no obligation to explain itself. In Sweden a municipality can veto a wind project at any point in the environmental permit process, does not have to justify the decision, and the decision cannot be appealed. Building a project pipeline in that environment is not an engineering skill, it is a decades-long relationship and site-selection skill, and it is the reason a 35-year-old developer with roughly 100 employees has a 14.8 GW pipeline while well-capitalised new entrants do not.

The second capability is transaction structuring. Look at how differently the recent deals were shaped. Pienava in Latvia went to a state utility, Latvenergo. Fageråsen (189 MW) was sold mid-development jointly with co-developer Dala Vind to OX2, with Eolus retaining a milestone claim that only paid out when OX2's own investment committee took final investment decision in February 2026. Roccasecca in Nevada (127 MW / 506 MWh) went all-cash to DESRI for approximately USD 66.9 million including cost reimbursement, with roughly USD 290 million of senior secured construction facilities arranged alongside (pv magazine USA, 1 April 2026). Centennial Flats was sold pre-construction on a long milestone tail. Stor-Skälsjön was sold as a 75% equity stake to MEAG with co-developer Hydro Rein retaining 25%. Each shape allocates construction risk, price risk and timing risk differently, and choosing the shape is where a developer either captures or gives away value.

Why it is a separate segment. Its economics are the opposite of a normal business. Revenue arrives in irregular blocks determined by closing dates and construction milestones outside the company's control. Costs are steady: roughly 100 salaried people developing projects continuously. The result is that the segment reliably loses money in quarters with no closings and produces most of the year's profit in quarters with one. Q2 2026 had no completed transaction and no project qualifying for progressive revenue recognition, and the quarter was loss-making. Q1 2026 had both the Roccasecca closing and the Fageråsen final investment decision milestone, and was strongly profitable, with the Roccasecca sale alone contributing a positive earnings effect of SEK 295 million (Q1 2026 report).

Competitive position. Fragmented and getting less so. In Swedish onshore wind more than 50 developers operate regionally, including Arise, Eolus and Rabbalshede Kraft, typically assembling 200 to 400 MW clusters on the strength of local stakeholder networks (Mordor Intelligence, Sweden Wind Energy Market). Eolus wins on permitting track record, on being able to build as well as develop, and on having a US platform that most Nordic peers lack. It loses where balance sheet size decides the outcome: it cannot hold a 1 GW offshore project through a decade of consenting the way Vattenfall or Statkraft can, which is precisely why the offshore write-downs landed on Eolus and not on a state-backed utility.

How management frames it. As the growth engine and the source of value creation, with an explicit preference for not holding assets. The CEO's framing in the Q1 2026 report was that after the Roccasecca sale "Eolus is back in an asset-light and liquid position" (Q1 2026 report, 6 May 2026).

Revenue mix. Project Development is the overwhelming majority of group sales in any year with transactions, and in a quarter without one it can be close to zero. This is the segment that determines whether the year is good or bad.

2.2 Asset Management

What it does. Technical and administrative management of operating renewable generation and storage assets on behalf of their owners: production monitoring, availability management, service contract administration, balancing and settlement, regulatory reporting, and commercial administration. The company's framing is "a full range of asset management services for carefree ownership" (company site).

Scale: 1,271 MW under management at 30 June 2026, 1,274 MW at 31 December 2025, 967 MW at 31 December 2024, 939 MW at Q1 2024. Roughly a third of growth in one year.

The core capability. Three things, none of which sound impressive individually. First, the installed base of relationships: Eolus manages assets for Aquila Capital, ewz, Munich Re, KGAL, Cubico Sustainable Investments and Mirova, and those mandates came attached to sales it made. Second, jurisdictional operating knowledge across six countries with different balancing markets, grid codes and tax treatments. Third, and most valuable, the fact that Eolus built many of these parks itself and therefore knows their as-built condition, their service contract history and their underperforming turbines in a way a third-party manager taking over cold does not.

Why it exists as a separate segment. Because its economics are the inverse of development. Revenue is contracted, recurring and predictable, typically on multi-year agreements (the Mirova mandate runs fifteen years). It does not consume development capital. Its purpose inside the group is to make a violently cyclical business partially fundable: in the Q2 2026 report, with Project Development contributing almost nothing, management specifically noted that Asset Management contributed stable earnings.

It is also a customer-retention device. A buyer who has Eolus managing three of its parks is a warm buyer for the fourth. The Mirova structure makes this explicit: Eolus sold the parks, kept the fifteen-year management contract, and agreed to co-develop batteries at the same sites. One sale produced three separate ongoing commercial relationships.

Competitive position. This is a more commoditised market than development. Competitors include the asset management arms of other developers (OX2 manages 38 wind parks totalling 2.1 GW), independent technical asset managers, and the original equipment manufacturers themselves, since Vestas, Siemens Gamesa and Nordex all sell long-term service agreements and would happily extend into commercial management. Eolus wins on incumbency and on being the party that built the asset. It loses on scale: a manager with 5 GW under contract has better spare-parts leverage and better data than one with 1.3 GW.

How management frames it. As the stabiliser, not the growth story. It is consistently described in the quarterly reports as delivering stable earnings, and its MW growth is presented as an output of development success rather than as a standalone target.

Revenue mix. Small in absolute terms relative to a transaction year, but positive and non-volatile. In a quarter like Q2 2026 it is most of what the group actually earned from operations.

Segment comparison

3. Products and business detail

3.1 The product catalogue, by technology

Onshore wind (5,834 MW in portfolio, June 2026). The founding product and still the largest single line. Eolus has built more than 800 turbines. Modern Eolus projects are large-rotor, tall-tower machines optimised for moderate Nordic wind speeds: Fågelås, completed in 2025, uses Vestas turbines standing 250 metres to blade tip, the tallest Eolus has built. Geography matters enormously to the economics here, because Sweden is divided into four electricity price areas and the northern ones (SE1, SE2) have suffered persistently low and occasionally near-zero prices. Fågelås, Dållebo and Boarp all sit in SE3, the central zone, which is materially better ground.

Utility-scale solar (4,988 MW). Predominantly the US and Poland. In the US, western states: California, Arizona, Nevada, Idaho, Utah, Wyoming, Washington and New Mexico. In Poland, where Eolus has been active since 2021, most of the portfolio is solar with some early-phase onshore wind. Solar development is faster than wind (shorter measurement campaigns, less contentious permitting) but far more exposed to grid interconnection queues and, in the US since July 2025, to federal tax credit deadlines.

Battery energy storage (2,975 MW). The fastest-growing line and, on recent evidence, the most commercially successful. Eolus has now sold at least four US storage projects: Cald BESS to Aypa Power in 2021, Centennial Flats (as the storage half of a hybrid) to Copia Power in 2022, Pome (100 MW / 400 MWh, Poway, California) with an enterprise value in the USD 230 to 235.5 million range announced January 2025, and Roccasecca (127 MW / 506 MWh, Boulder City, Nevada) to DESRI in February 2026. The Nordic storage business is younger but is being seeded directly off the existing wind estate: the Mirova agreement covers joint battery development at Fågelås, Dållebo and Boarp.

Offshore wind (1,000 MW). Now effectively a stranded line item in Sweden. Two projects of roughly 1,000 MW each, Västvind and Najaderna, were rejected on 16 July 2026. Approximately two thirds of the roughly SEK 240 million of impairments taken in Q4 2025 already related to offshore wind, following a reassessment of Swedish and Finnish offshore market conditions (Eolus release, 27 January 2026).

Hybrid projects. Co-located generation plus storage. Centennial Flats is the template: 500 MW of solar next to 267 MW of batteries, expected to produce around 1 TWh a year. Hybrids are increasingly the default in the US and are appearing in Finland and Poland, because batteries monetise curtailment and capture intraday spreads that a bare solar plant gives away.

Asset management and technical advisory services. Sold as multi-year contracts, usually attached to a divestment.

3.2 How a project is actually made

The development sequence, roughly, and it is worth understanding because every risk in Section 8 sits somewhere on it:

1. Site origination. Identify land with wind or irradiation resource, plausible grid proximity, and a municipality that will not veto. In Sweden that last filter has become the binding one.
2. Land control. Lease options from landowners. In Nordic forestry country that can mean dozens of private owners plus state forest holdings.
3. Resource measurement. Wind measurement campaigns run for a year or more. Solar needs far less.
4. Grid connection. Apply and queue. In the US this is now the single longest pole in most solar and storage projects.
5. Permitting. Environmental impact assessment, county or municipal consent, national approval where the project sits in an exclusive economic zone (offshore) or crosses defence interests.
6. Procurement. Turbine or module and battery supply agreements. Eolus has a long relationship with Vestas as a supplier.
7. Offtake. A power purchase agreement or tolling agreement with a creditworthy counterparty, which is what converts the project from a merchant power gamble into a bankable asset.
8. Financing and sale. Either sell the project pre-construction (Centennial Flats), sell during construction (Roccasecca), or build and sell at or after completion (Fågelås, Dållebo, Boarp, which began construction in spring 2024, reached substantial completion in Q3 2025 and were sold in December 2025).
9. Construction management. Eolus can act as construction manager even on assets it has already sold, which is how it continues to earn on Centennial Flats four years after the sale.
10. Asset management. The fifteen-year tail.

Note step 8 carefully. The decision of when in the lifecycle to sell is the central commercial judgement of this business, and it moves both price and risk. Selling early, as with the mid-stage US solar and battery project sold on 4 September 2026 for up to USD 6.5 million (USD 2.0 million upfront, USD 4.5 million contingent on milestones), realises a modest sum with no construction exposure (Eolus release, 4 September 2026). Selling late, as with the Mirova parks, realises far more per MW but requires Eolus to carry construction capital and completion risk on its own balance sheet first, which is exactly what took the company into net debt during 2025 and out again in Q4.

3.3 Geographies

Sweden. Home market, majority of offices and employees, and roughly half the project portfolio. Full technology range: wind, solar, batteries, hybrids. This is also the market with the worst current permitting environment, discussed in Section 6.

United States (since 2015). Now a large share of the portfolio and the source of the largest cash events. Eolus North America operates in the western states. The US business has produced four battery divestments and one large hybrid. It is also where the policy risk is now most acute, following the July 2025 federal tax legislation.

Finland (since 2014). Country-managed local organisation. Focus on onshore wind, battery storage and hybrids.

Poland (since 2021). Mostly solar, some early-stage onshore wind, plus storage and hybrids. The newest organisation of any scale.

Baltics (Estonia since 2006, Latvia since 2011). Turbines built in Estonia from 2006. The recent centre of gravity is Latvia, where Eolus developed Pienava, described as the largest onshore wind project in Latvia to date, and sold it to Latvenergo in June 2025.

Norway (since 2012). Two completed projects: Stigafjellet (2020) and Øyfjellet (2023). Øyfjellet is the one with the fifteen-year, 330 MW Alcoa power purchase agreement covering roughly 1.2 TWh a year.

3.4 Milestones that changed the business

• 1991: First installation, two Vestas 225 kW turbines, Grönhögen, Öland.
• 2004: 100th turbine; head office moves to Hässleholm.
• 2014 to 2015: The Google and ewz transactions. Four small Swedish parks totalling 29 turbines and roughly 60 MW, with a ten-year Google power purchase agreement supplying its Hamina data centre in Finland, sold to Zurich's municipal utility ewz for EUR 95 million with Aquila Capital as co-investor, alongside a twelve-year technical services agreement and a four-year administrative agreement. This is the first full expression of the modern model: build, contract with a corporate offtaker, sell to a foreign institution, keep the service contract.
• 2015: US market entry.
• 2017: Kråktorpet and Nylandsbergen, 61 turbines, 232 MW, sold to Aquila Capital for a preliminary purchase price of approximately EUR 264 million. Proof the model scaled.
• 2021: First US battery sale (Cald BESS to Aypa Power). Poland entry.
• 2022: Centennial Flats sold to Copia Power, the first combined solar and storage divestment. Stor-Skälsjön, 260 MW, 75% sold to MEAG at an enterprise value of EUR 361.4 million on a 100% basis, with Eolus selling its full 51% for EUR 18.4 million.
• May 2025: First green bond. SEK 550 million of senior secured floating-rate green bonds at three-month STIBOR plus 750 basis points, maturing 2029, listed on Nasdaq Stockholm's sustainable bond list, under a framework rated Dark Green by S&P Global, with the framework permitting subsequent issuance up to SEK 1.5 billion in aggregate (Financing).
• 2025: Name change to Eolus AB. 448 MW brought into operation during the year.
• December 2025: Mirova transaction turns net debt into net cash.
• February 2026: Roccasecca all-cash sale to DESRI.
• May 2026: Bond tender offer at 104.0% of nominal, SEK 71.25 million repurchased, leaving SEK 478.75 million outstanding.
• July 2026: Both Swedish offshore projects rejected.

4. Customers

4.1 Who buys, and who inside the buyer decides

Institutional infrastructure investors and their asset managers. Named counterparties over time include Mirova (an affiliate of Natixis Investment Managers), MEAG (Munich Re's asset manager), Aquila Capital, KGAL, ewz (the City of Zurich's electricity utility), and Cubico Sustainable Investments. The buying decision here runs through an investment committee and then a technical and legal due diligence process typically lasting three to six months. The criteria are: is the permit final and unappealable, is the grid connection firm, is the offtake contract with a counterparty our credit team will accept, is the turbine supply and service agreement bankable, and what is the yield after we model production with a conservative loss assumption. A developer fails this process on documentation quality far more often than on price.

Independent power producers and strategic developers. DESRI bought Roccasecca. Copia Power (Carlyle) bought Centennial Flats. Aypa Power bought Cald. OX2 bought Fageråsen jointly from Eolus and Dala Vind. These buyers are technically sophisticated, will buy earlier in the lifecycle, and pay less per MW in exchange for taking more development and construction risk. The decision-maker is a development or M&A team, and the cycle can be faster.

State-owned utilities. Latvenergo bought Pienava. This class of buyer is often the natural owner of the largest asset in a small market and brings its own political weight to grid and permitting matters.

Corporate and municipal electricity buyers. Alcoa (fifteen years, 330 MW, roughly 1.2 TWh a year at Øyfjellet), Google (ten years, four Swedish parks supplying Hamina), Sonoma Clean Power (ten-year tolling on the Pome battery). The decision-maker here is an energy procurement function, sometimes with a sustainability mandate attached. The criteria are volume, price, shape, and whether the seller can actually deliver the asset on time. A corporate offtake signed before the sale is one of the most value-additive things a developer can do, because it converts merchant price exposure into a contracted cash flow the institutional buyer's credit committee can underwrite.

Existing asset owners, for management services. Same institutions as above, now buying a service rather than an asset.

4.2 Why they choose Eolus

Not because of price. Developed projects are competitively marketed and several of these transactions ran as processes. The reasons that come through the transaction record are more specific:

• Delivery record. The CEO's framing after the Mirova closing was that this was the fourth transaction in four consecutive quarters (Eolus release, 23 December 2025). For a buyer choosing between a developer that closes and one that renegotiates, that history is the product.
• Ability to build, not just permit. Eolus completed and handed over 448 MW during 2025. A buyer that wants a completed asset rather than a permit needs a counterparty that can manage construction, and most pure developers cannot.
• Permit quality in hard jurisdictions. A Swedish permit that has survived municipal consent is scarce, which is a direct function of Section 6.
• The service tail. A buyer who does not want to build an operations organisation in Sweden can buy the asset and the operator in one transaction.
• Offtake already attached. Fågelås, Dållebo and Boarp had a fifteen-year power purchase agreement for a significant share of production, signed 29 September 2025, before the December sale.

4.3 Switching costs

Low on the development side, high on the management side, and the asymmetry matters.

There is no lock-in on project sales. Every transaction is a fresh competition, and a buyer who bought a park in 2017 has no obligation to look at the 2026 one. This is the honest weakness in the business model: Eolus has to win each sale again.

Asset management is stickier. A fifteen-year contract on a park Eolus built, with Eolus holding the as-built records, the service-contract history and the local balancing arrangements, is genuinely inconvenient to move. Transitioning a technical management mandate means re-papering grid and balance responsibility, rebuilding SCADA integrations and losing institutional memory of the asset, for a fee saving measured in basis points of asset value. This is why the MW under management line has risen monotonically even through a bad period for the development business.

On the offtake side, switching costs are contractual rather than operational: a fifteen-year power purchase agreement is a fifteen-year power purchase agreement.

4.4 Concentration

Transaction concentration is extreme by construction and cannot be engineered away. In any given quarter, one or two counterparties account for essentially all of the development revenue, because that is how many deals close. In Q1 2026 it was DESRI and OX2. In Q4 2025 it was Mirova. In Q2 2026 it was nobody, and that is exactly what the near-zero sales figure was saying.

This is not the same thing as customer dependency. Across the last decade the buyer list is genuinely diverse (Swiss municipal utility, German reinsurer's asset manager, French bank's sustainability affiliate, Latvian state utility, two US private-equity-backed platforms, a Swedish peer developer). No single buyer is structurally necessary. What Eolus is dependent on is the existence of a functioning market for de-risked renewable assets, not on any one participant in it.

Asset management concentration is more real: 1,271 MW spread across a modest number of institutional owners means losing one large mandate would be visible.

4.5 Contract structures and what they mean for predictability

Four distinct structures, in rising order of how much they smooth results:

1. Pre-construction sale with milestone tail. Centennial Flats. Small upfront, payments over years as construction and commercial operation milestones are hit. Produces a long, partly predictable cash tail but means the developer is exposed to a third party's construction schedule for years. Eolus has been collecting on a 2022 sale throughout 2026.
2. Mid-development sale with contingent consideration. The September 2026 US solar and battery sale (USD 2.0 million upfront of up to USD 6.5 million) and Fageråsen, where the milestone only paid when OX2 took final investment decision. Low capital at risk, low and uncertain proceeds.
3. All-cash sale on or near completion. Roccasecca (approximately USD 66.9 million, all cash, including cost reimbursement). Cleanest outcome, highest proceeds, but Eolus had to fund development to that point.
4. Build-then-sell with retained services. Fågelås, Dållebo, Boarp. Highest value capture per MW, largest working capital swing, and the one that generates the fifteen-year fee tail.

The consequence for anyone reading the quarterly numbers: Eolus's revenue recognition is partly progressive (projects under construction can qualify for percentage-of-completion recognition) and partly event-driven. In Q2 2026 neither channel produced anything, which is why sales were nominal. That is a timing statement, not a demand statement.

5. Competitive landscape

5.1 The structural fact that dominates everything else: the listed Nordic peer group has been bought

Two years ago an investor comparing Nordic listed renewable developers had a choice of OX2, Arise, Cloudberry and Eolus. Today:

• OX2 was acquired by Otello BidCo, an EQT vehicle, at SEK 60 per share, valuing it at roughly SEK 16.35 billion. Otello reached 98.81% of shares and votes, and OX2's last trading day on Nasdaq Stockholm was 21 October 2024 (EQT; MarketScreener).
• Arise AB received a proposal from Aneo Holding AS on 26 November 2025 at approximately SEK 1.8 billion, and Nasdaq approved its delisting with a last trading day of 26 January 2026.
• Cloudberry Clean Energy ASA remains listed in Oslo.
• Eolus remains listed in Stockholm.

So the competitive set is now mostly private capital, state utilities and large European energy groups. That cuts both ways for Eolus, and honestly. On one side, private owners of OX2 can outspend a listed developer that must show quarterly results, and EQT's ownership removes OX2's need to divest on a public-market cadence. On the other, two competitors being taken out at premium valuations is evidence that the asset class of "permitted Nordic pipeline plus development team" has a real price to strategic and financial buyers.

5.2 Named competitors

In Nordic onshore wind and hybrid development: OX2 (now EQT-owned, 11 European markets plus Australia, 38 wind parks and 2.1 GW under management), Cloudberry Clean Energy, Rabbalshede Kraft, Statkraft, Vattenfall, Fortum, and the development arms of RWE and Ørsted. Mordor Intelligence counts more than 50 developers active in Swedish onshore wind, which tells you the barrier to entering is low even though the barrier to completing is high.

In Swedish offshore wind: a much more consolidated field, where Vattenfall, OX2 and Statkraft hold the largest lease positions and the strongest relationships with grid and maritime regulators. This is precisely the field Eolus has now exited in Sweden by government decision.

In US utility-scale solar and storage development: Invenergy, Intersect Power, Arevon, Pattern Energy and the development arms of NextEra, plus the buy-side platforms that sometimes also develop (DESRI, Copia Power, Aypa Power). Eolus is a small participant here by any measure, and it competes by originating in less-contested western-state locations and selling early rather than by out-financing anyone.

In asset management: OX2, independent technical asset managers, and the original equipment manufacturers' own service organisations (Vestas, Siemens Gamesa, Nordex), who already hold the long-term service agreements on much of the installed base and could extend upward into commercial management.

5.3 Where Eolus wins and where it loses

Wins:

• Permitting and local-relationship depth in Sweden, built over 35 years. Roughly 13% of Sweden's turbines is not a marketing claim that a new entrant can replicate with capital.
• Willingness and ability to build, which widens the buyer universe to institutions that want a completed asset.
• Technology breadth relative to its size: a genuinely material solar and storage pipeline, not a wind company with a solar slide.
• A US platform with four completed storage divestments, which most Nordic peers of this size do not have.
• Balance sheet position entering the current downturn. The Q4 2025 transactions generated net cash flow of just over SEK 500 million and turned net debt into net cash, and the equity-to-assets ratio was 65% at 30 June 2026 against a target floor of 30%.

Loses:

• Cannot carry very long-dated, capital-hungry projects. Offshore wind is the proof: two 1 GW projects, years of capitalised cost, terminated by ministerial decision, with the loss falling on a company of Eolus's size rather than on a national utility's balance sheet.
• No cost advantage anywhere. It buys turbines from the same suppliers at worse volumes than Vattenfall or RWE.
• Must re-win every project sale. There is no recurring development revenue.
• Small relative to the US competitive set, where scale in interconnection queue position, tax equity relationships and equipment procurement matters a great deal.

5.4 Barriers to entry

Genuinely mixed, and it is worth being precise rather than asserting a moat.

Low barriers: anyone can lease land and file a permit application. More than 50 developers do it in Sweden alone. Capital is available to credible teams.

High barriers: completing. The scarce resources are (a) municipal consent in Sweden, which is discretionary, unexplained and unappealable, (b) grid connection capacity, which is queued and finite, and (c) a delivery record long enough that an institutional buyer's investment committee will sign. Item (c) is the one that compounds, and it is the closest thing Eolus has to a durable advantage: a first-time developer with a good permit still has to convince MEAG or Mirova to underwrite its construction management.

Weakening barrier: the value of a permit is only as good as the economics of building on it. If Nordic power prices stay low enough that fully permitted projects do not reach final investment decision (536 MW of fully permitted Swedish projects were awaiting investment decisions, per Swedish Wind Energy Association data), then a permitting advantage converts into inventory rather than into cash.

5.5 Structural shifts in progress

1. Consolidation of the listed peer set into private and strategic hands (OX2 to EQT, Arise to Aneo). Eolus is now one of very few ways to own this business model in public markets.
2. Technology rotation from wind to solar-plus-storage. Eolus's own portfolio mix says it: storage grew from 2,975 MW in June 2026 versus a solar line that shrank, while onshore wind fell from 6,409 MW at end-2025 to 5,834 MW. Some of that is projects leaving the pipeline via sale, and some is pruning.
3. Offshore wind in Sweden moving from open application to tender. Following the second wave of rejections in July 2026, the government is shifting to a tender-based allocation system. That resets the offshore competitive field entirely and favours participants who can wait.
4. Nordic price weakness pushing value from generation towards flexibility. In 2025 the mFRR (manual frequency restoration reserve) market became the dominant revenue stream for grid-scale batteries across the Nordics, and battery co-location is emerging as a hedge against curtailment. Eolus co-developing batteries at the three Mirova sites is a direct response.

5.6 Competitor comparison

Market caps are peer-size references only, at the dates shown. Where I could not verify a figure in this research I have said so rather than estimate.

6. Industry

6.1 What drives demand

The structural case Eolus itself points to is demand growth: electricity consumption in Europe and the United States is expected to more than double by 2050, while renewable generation is cost-competitive (Eolus investor relations). Electrification of transport, industrial heat, heavy industry (aluminium, steel, hydrogen) and data centres are the sources.

Underneath that long arc, four nearer-term drivers determine whether a developer sells projects in any given year:

1. The forward price of electricity in the relevant price area. Low forward prices stop investment decisions dead, permit or no permit.
2. The availability of a creditworthy offtaker. Corporate power purchase agreements are what make merchant projects bankable. Google, Alcoa and Sonoma Clean Power on the Eolus record are examples of this working.
3. The cost of capital for infrastructure buyers. Rising rates in 2022 to 2024 widened required yields and compressed what buyers would pay per MW. Eolus's own commentary in September 2024 cited normalising costs and declining interest rates as supportive of its US outlook.
4. Policy and subsidy mechanics, which in the US changed abruptly in July 2025 (below).

6.2 Size and growth

Swedish wind. Swedish wind power produced approximately 40.8 TWh in 2024, the highest on record. The Swedish Wind Energy Association projected cumulative capacity of around 19.5 GW by the end of 2026, with longer-term third-party projections of roughly 30 GW by 2030.

But new build is collapsing. Turbines totalling 446 MW were ordered in Sweden during 2024. 536 MW of fully permitted projects were sitting without an investment decision, awaiting better market conditions. EnergyWatch reported no new Swedish wind investment decisions in a recent fourth quarter and a forecast of roughly 280 MW for 2027. The Swedish press framing has been blunt: interest in building new wind power is dead (Sweden Herald). The stated cause is low electricity prices, particularly in northern Sweden, linked to electrification proceeding more slowly than the generation build-out assumed. Prices in parts of the Nordic market have been expected to sit close to zero for extended periods between 2025 and 2027.

Nordic energy storage. The counterweight. The Nordic energy storage market was valued at approximately USD 4.35 billion in 2024 and is projected to reach USD 18.41 billion by 2035, a compound annual growth rate of about 13.7%, with battery energy storage systems accounting for roughly 40 to 50% of total market value (DataNext Research). In 2025 the mFRR market became the dominant revenue stream for grid-scale batteries across the region.

Put those two facts next to each other and you have the single most important industry sentence for Eolus: its oldest and largest product line is in a domestic investment freeze while its newest one is compounding at double digits. That is precisely why the portfolio mix has been rotating and why batteries are being co-developed on existing wind sites.

6.3 Where Eolus sits in the value chain

Upstream of Eolus: turbine manufacturers (Vestas, Siemens Gamesa, Nordex), module and battery suppliers, EPC contractors, and grid operators who control connection. Eolus is a price-taker on all of these.

Downstream of Eolus: institutional asset owners and independent power producers who buy the finished asset, and the electricity buyers who contract its output.

Eolus occupies the origination-to-commissioning slice, plus a service slice after. It takes no long-term commodity price risk by design, which is both its defence (a bad decade for power prices does not impair an owned fleet it does not have) and its vulnerability (it has no contracted cash flow to fall back on when buyers stop buying).

6.4 Regulation, and the two policy shocks that matter

Sweden: the municipal veto. In Sweden the municipality hosting a wind project can veto it at any point in the environmental permit process, does not have to give reasons, and the decision cannot be appealed. Historically municipalities approved somewhere between 50% and 70% of projects. In 2025 that collapsed. Between 1 January and 12 September 2025, 26 of 29 proposed onshore projects were stopped, covering 346 of 359 turbines and a potential 7.3 TWh (Baltic Wind). Windpower Monthly reported the veto rate at around 90%. The government's response is financial: incentive payments to municipalities of SEK 340 million in 2025, an estimated SEK 370 million in 2026 and SEK 400 million in 2027. Whether money changes a political calculation is an open question.

Sweden: offshore and national defence. This is the shock that hit Eolus directly. On 16 July 2026 the Swedish government approved two offshore projects (Fyrskeppet and Vidar) and rejected eleven, including Eolus's Västvind and Najaderna, on the grounds that the Armed Forces had shown the turbines would impair Sweden's ability to detect incoming missiles and submarine movements (Windtech International; offshoreWIND.biz). Across two years Sweden has now rejected 24 offshore wind farms on defence grounds. The government is moving offshore allocation from an open application process to a tender system.

United States: the One Big Beautiful Bill Act. Signed 4 July 2025, this rewrote the incentive structure the US pipeline was built on. Wind and solar facilities lose the section 48E investment tax credit and section 45Y production tax credit unless construction begins within one year of enactment (by 4 July 2026) or the facility is placed in service by 31 December 2027. Standalone energy storage keeps those credits with a phase-down starting only in 2034. New Foreign Entity of Concern restrictions apply to project ownership, supply chains and tax credit buyers, with recapture provisions that can claw back previously claimed credits if a project later becomes non-compliant. An executive order of 7 July 2025 directed Treasury to tighten the begun-construction safe harbours, potentially requiring that a substantial portion of a facility actually be built (Latham & Watkins; Kirkland & Ellis).

The asymmetry in that legislation is the single most useful thing to know about Eolus's US business. Solar and wind got a cliff in 2026 and 2027. Storage got until 2034. Eolus's US divestment record over the last five years is dominated by storage: Cald, Pome, Roccasecca, and the storage half of Centennial Flats. The part of the US pipeline with the most favourable policy runway is the part Eolus has been most successful at selling.

6.5 Cyclicality

This industry is cyclical on three separate clocks, and they do not synchronise:

• The interest rate clock, which sets the discount rate infrastructure buyers apply and therefore the price per MW. Rising rates in 2022 to 2024 compressed transaction values; the subsequent easing helped.
• The power price clock, driven by hydrology, nuclear availability, gas prices and the pace of industrial electrification. Nordic prices have been weak and are expected to remain so into 2027.
• The policy clock, which moves in discrete jumps on election and legislative timetables rather than smoothly.

For a developer, the worst configuration is what 2025 and 2026 delivered: weak power prices freezing domestic investment decisions, a hostile permitting environment closing the funnel, and a policy reset in the growth market. The current cycle position is therefore genuinely bad, which is also why the company's cash position matters more than usual.

6.6 Tailwinds and headwinds at industry level

Tailwinds: electricity demand growth from electrification and data centres; falling battery costs; the rise of Nordic ancillary service markets as a battery revenue stream; renewables being cost-competitive on a new-build basis; consolidation demonstrating institutional appetite for development platforms.

Headwinds: near-zero Nordic power prices for extended periods to 2027; Swedish municipal veto rates around 90%; Swedish offshore effectively closed pending a new tender regime with defence screening; US tax credit cliffs for wind and solar in 2026 and 2027 plus Foreign Entity of Concern compliance uncertainty; grid connection queues; slower-than-assumed industrial electrification in northern Sweden.

7. Growth triggers

Sourced from the six most recent reporting events and the dated regulatory releases issued alongside them. As noted at the top, Eolus does not publish verbatim webcast transcripts, so each item below is attributed to the interim report or regulatory release published with that quarter's webcast rather than to audio.

• Centennial Flats phase three commercial operation, expected before end-2026, releasing approximately USD 12 million of final gross proceeds. Repeated and upgraded across multiple quarters; the forecast for remaining proceeds was raised in August 2026. (Q2 2026 report, 27 August 2026; release of 14 August 2026)

  "The net impact on Eolus's EBIT and cash flow for Q3 2026 is expected to be approximately USD 20 million."

  • Eolus release, 14 August 2026

• Battery storage development at the three former Mirova sites (Fågelås, Dållebo, Boarp), to be developed jointly with the new owner. A new pipeline created out of a completed divestment, on land and grid connections already secured. (Q4 2025 / Year-End Report, 11 February 2026; originating release 23 December 2025)

• Roccasecca reaching commercial operation during 2026, completing the 127 MW / 506 MWh Nevada battery project sold to DESRI and closing out Eolus's remaining obligations on it. (Q1 2026 report, 6 May 2026)

• Fageråsen construction proceeding under OX2 ownership following the customer's final investment decision on 189 MW and 27 turbines in February 2026, which triggered the milestone payment to Eolus. Repeated across the Q4 2025 and Q1 2026 reports. (Q1 2026 report, 6 May 2026)

• Contingent consideration of up to USD 4.5 million on the mid-stage US solar and battery project sold on 4 September 2026, payable on achieved project milestones, on top of the USD 2.0 million received upfront. (Release of 4 September 2026, ahead of the Q3 2026 report)

• Asset management portfolio continuing to grow off completed divestments, having moved from 967 MW at end-2024 to 1,274 MW at end-2025 and 1,271 MW at 30 June 2026, with the fifteen-year Mirova mandate now embedded. (Q2 2026 report, 27 August 2026; Year-End Report 2025, 11 February 2026)

• Deployment of the post-transaction cash position into the existing 14.8 GW pipeline in a market where competitors are capital-constrained. Management has stated this repeatedly across the last three reports.

  "Eolus is back in an asset-light and liquid position, ready to take on both challenges"

  • CEO Per Witalisson, Q1 2026 report, 6 May 2026

  "As a pure-play developer with a strong financial position, Eolus is well equipped to navigate a hesitant market."

  • CEO Per Witalisson, Q2 2026 report, 27 August 2026

• 282 MW under construction at 31 December 2025, the near-term source of progressive revenue recognition and of the next completion-and-handover events. (Year-End Report 2025, 11 August reporting date 11 February 2026)

Trigger summary

8. Key risks

8.1 Political and regulatory termination risk on permitted or near-permitted projects. High probability, demonstrated severity.

This is not a theoretical risk for Eolus; it has now materialised twice in nine months. The mechanism is that development costs are capitalised on the balance sheet as a project progresses, and a government or municipal decision can extinguish the entire asset with no appeal and no compensation.

In Q4 2025 the board recognised impairments of approximately SEK 240 million after reassessing offshore wind market conditions in Sweden and Finland and reviewing the total portfolio, with about two thirds relating to offshore (release, 27 January 2026). Then on 16 July 2026 the Swedish government rejected Västvind and Najaderna outright on defence grounds, and the remaining approximately SEK 48 million of Västvind capitalised cost on the 30 June 2026 balance sheet was written down.

Onshore, the same mechanism operates through the municipal veto, which requires no justification and cannot be appealed, and which rejected roughly 90% of Swedish projects in 2025.

The calibration: this is a high-probability, recurring, moderate-to-large drag rather than a single catastrophic event, because the portfolio is diversified across six countries and four technologies. Management's own framing of that diversification is the mitigant:

"project portfolio with a size and risk diversification grants us stay power"

• CEO Per Witalisson, Q3 2025 report, 19 November 2025

8.2 The Swedish investment freeze turning the permitted pipeline into unsellable inventory. High probability, high impact.

A permit is only worth something if someone will build on it. With Nordic prices expected close to zero for extended periods to 2027, 536 MW of fully permitted Swedish projects were sitting without investment decisions, and the 2027 Swedish build forecast is roughly 280 MW. If buyers will not underwrite a project at any price Eolus finds acceptable, then development spend accumulates on the balance sheet, no transaction closes, and quarters look like Q2 2026: near-zero sales and an operating loss carried by a fixed cost base of around 100 employees.

This is the risk that shows up in the numbers first and in the headlines last. Eolus explicitly attributed the Q2 2026 sales collapse to no completed transactions and no projects qualifying for progressive revenue recognition. One such quarter is timing. Four in a row would be a business model problem.

8.3 US policy cliff on the solar and wind half of the US pipeline. Moderate-to-high probability, moderate impact.

Under the July 2025 legislation, US wind and solar projects lose the investment and production tax credits unless construction began by 4 July 2026 or the facility is placed in service by 31 December 2027. The 7 July 2025 executive order directed Treasury to tighten the begun-construction safe harbours. Separately, Foreign Entity of Concern rules now restrict ownership, supply chains and tax credit buyers, with recapture if a project later becomes non-compliant.

The mechanism for Eolus: a US solar project that cannot credibly claim begun construction is worth materially less to a buyer, because the buyer's returns were underwritten on a tax credit that may not exist. Eolus's solar line was 4,988 MW at June 2026, so the exposure is real in MW terms even though Eolus's realised US proceeds have come mostly from storage, which keeps its credits until a phase-down beginning in 2034. Note also that a Swedish company developing in the US is exactly the sort of party that has to prove Foreign Entity of Concern compliance rather than assume it, which adds diligence friction to every US sale.

8.4 Bond covenants converting a bad year into a capital-return prohibition. Already realised.

The senior secured green bonds (2025/2029, ISIN SE0024320774) contractually prevent dividend payments and share repurchases following a negative annual result. After the loss-making 2025, Eolus stated plainly that the 2025 result prevented a dividend in 2026, and that executing share buybacks would require bondholder approval to amend the bond terms (Update regarding Eolus's capital allocation 2026).

The company was explicit that this reflects the covenant rather than a weak balance sheet, and the 65% equity-to-assets ratio at 30 June 2026 against a 30% target floor supports that. But the mechanism is worth understanding: in a business with structurally lumpy earnings, a covenant keyed to annual profit means one impairment year suspends shareholder distributions entirely, regardless of cash. The bonds carry three-month STIBOR plus 750 basis points, which is a high coupon and tells you what the debt market thought of this credit in May 2025.

8.5 Working capital and completion risk when Eolus builds before selling. Moderate probability, moderate impact.

The build-then-sell structure (Fågelås, Dållebo, Boarp) captures the most value per MW but requires Eolus to fund construction. During 2025 that pushed the group into net debt; the December 2025 Mirova closing generated just over SEK 500 million of net cash flow in the quarter and reversed it. That round trip worked. A construction cost overrun, a turbine delivery delay, or a buyer walking away between substantial completion and closing would not.

8.6 Concentration of the forward cash story in a small number of milestone events. Moderate probability, high impact on any single quarter.

Look at what is actually carrying the near-term outlook: Centennial Flats phase three, Roccasecca commercial operation, and up to USD 4.5 million of contingent consideration on a mid-stage US project. These depend on third parties' construction schedules. Centennial Flats has been paying since 2022 and phase three is still pending. A slip of one quarter in a project Eolus no longer controls moves Eolus's reported result by a large amount.

8.7 Loss of a large asset management mandate. Low probability, moderate impact.

The 1,271 MW under management is the only non-lumpy earnings in the group and is concentrated among a modest number of institutional owners. Contracts are multi-year and switching is inconvenient, so probability is low, but the loss of the single largest mandate would remove part of the only stable earnings line at exactly the moment the development business is not producing.

8.8 Credibility cost of withdrawn targets. Already realised, with a forward consequence.

On 27 January 2026 the board withdrew the financial target of at least SEK 1,400 million in total operating profit for 2025 to 2027, while keeping the return on equity target above 15% per year, the equity ratio target above 30% and the dividend policy of 20 to 50% of profit after tax. Withdrawing a three-year profit target in the first month of the target's second year is a significant admission. The forward consequence is that investors now have no multi-year quantitative commitment to hold management to, which makes the qualitative commitments in Section 9 the only available yardstick.

9. Walk the talk

The six reporting events: Q1 2025 (May 2025), Q2 2025 (26 August 2025), Q3 2025 (19 November 2025), Q4 / Year-End 2025 (11 February 2026), Q1 2026 (6 May 2026), Q2 2026 (27 August 2026). The most recent is 23 days old.

A necessary caveat before the assessment: because Eolus does not publish webcast transcripts and the third-party archives are subscriber-gated, what follows tracks management's written commitments in the interim reports and dated regulatory releases against subsequent outcomes. That is a narrower evidence base than a full transcript record, and I will not claim consistency or inconsistency on statements I could not read.

The arc

Q1 2025 set a confident tone on transaction execution. The quarter carried the closing of the Pome battery sale in California at an enterprise value in the USD 230 to 235.5 million range, with a SEK 275 million milestone payment received in February, and the completion and handover of Stor-Skälsjön, co-developed with Hydro Rein, in March. The CEO described a strong quarter while noting adjustments to offshore wind and financial market developments. It is worth marking that offshore caution appears this early, in the first of the six reports, nine months before the impairment.

Q2 2025 made the claim that would define the year: Eolus can transact in a bad market. Having sold Pienava to Latvenergo in June (described as the largest onshore wind project in Latvia to date) and Fageråsen to OX2 jointly with Dala Vind in July, the CEO wrote:

"The fact that Eolus has completed two transactions over a short period of time in a challenging market is strong proof of this."

• Q2 2025 report, 26 August 2025

And, on organisational capability:

"I am confident in my colleagues' ability to act with bravery and determination to achieve excellent results and value creation also in a weaker market."

• Q2 2025 report, 26 August 2025

This was also the quarter of the green bond issue (SEK 550 million) and the refinancing of existing debt, and of the name change from Eolus Vind AB to Eolus AB. The equity-to-assets ratio improved from 44% a year earlier to 59%, and asset management grew from 967 MW to 1,186 MW.

Q3 2025 was a loss-making quarter in which management leaned on portfolio breadth rather than promising a recovery. The quarter's sales were mainly revenue recognition from the US battery project Pome. The green bonds began trading on Nasdaq Stockholm on 2 July, the Fageråsen sale to OX2 closed on 25 July, and on 29 September Eolus signed a fifteen-year power purchase agreement for a significant share of production at Fågelås, Dållebo and Boarp. The CEO's framing was defensive and, in hindsight, accurate: portfolio size and risk diversification grant staying power. Notably, no promise of a Q4 recovery was made in that language, and yet a Q4 recovery arrived.

Q4 and Year-End 2025 is where the promise-keeping and the promise-breaking sit side by side, and the juxtaposition is the most informative thing in the six reports.

Kept: on 23 December 2025 Eolus completed the sale of Fågelås, Dållebo and Boarp (88 MW, SE3) to Mirova, with a fifteen-year asset management agreement retained and joint battery development agreed at all three sites. The CEO's own framing:

"This deal is an excellent example of Eolus's capacity to execute projects and deliver market-leading assets to investors"

• Eolus release, 23 December 2025

He described it as the fourth transaction in four consecutive quarters, which is a checkable claim and is consistent with the record: Pome (Q1 2025), Pienava (Q2 2025), Fageråsen (Q3 2025), Mirova (Q4 2025). The balance sheet effect was delivered as described:

"The fourth quarter's transactions generated a net cash flow of just over SEK 500 million, letting us turn net debt into net cash."

• CEO Per Witalisson, Year-End Report 2025, 11 February 2026

Broken, in the same window: on 27 January 2026, two weeks before the year-end report, the board pre-announced approximately SEK 240 million of impairments (about two thirds offshore wind) and withdrew the financial target of at least SEK 1,400 million of total operating profit for 2025 to 2027. The other three targets, return on equity above 15%, equity ratio above 30% and the 20 to 50% dividend policy, were kept.

Two things deserve credit here even though the outcome was bad. First, the board pre-announced the impairment and the target withdrawal on 27 January rather than burying it in the 11 February report, which is the harder and more transparent sequence. Second, the offshore caution had been flagged in the CEO's own commentary as far back as Q1 2025. This was not a surprise dressed up as one.

What deserves criticism: a three-year operating profit target abandoned thirteen months into the three years is a target that was set with insufficient regard for the permitting risk the company itself had been describing.

Q1 2026 delivered on the recovery narrative. The Roccasecca battery project (127 MW) was sold to DESRI for approximately USD 66.9 million, with a stated positive earnings effect of SEK 295 million, and a customer's final investment decision on Fageråsen triggered a milestone payment. The equity-to-assets ratio stood at 60% and 1,271 MW was under management. And the board did exactly what the bond terms required and what it had signalled: no dividend for 2025.

Q2 2026 was an honest bad quarter. No completed transactions, no projects qualifying for progressive revenue recognition, an operating loss, and a CEO statement that does not attempt to spin it:

"As a pure-play developer with a strong financial position, Eolus is well equipped to navigate a hesitant market."

• Q2 2026 report, 27 August 2026

"Hesitant market" is the correct word for a market where 90% of Swedish onshore projects were vetoed and 2027 build is forecast at 280 MW. The report also disclosed the Västvind write-down following the 16 July rejection, and disclosed the Centennial Flats milestone receipt with an explicit quantified forward guide for Q3 2026 of approximately USD 20 million net.

Promise versus outcome

Assessment

This is management that delivers transactions and does not deliver forecasts.

On the operational commitments, the record is genuinely good and unusually checkable. Five divestments in six quarters, in what management itself called a hesitant market, with the balance sheet moving from net debt to net cash exactly as described. The asset management line has grown every single year. The Centennial Flats tail, structured in 2022, has now paid USD 110.0 million of a USD 116.9 million expectation, which is a long-dated commitment substantially honoured. When Eolus says it will sell a project, it sells the project.

On the forward-looking quantitative commitments, the record is poor. A three-year profit target abandoned in month thirteen is the clearest possible evidence that the target was built on permitting assumptions the company's own commentary already doubted. The Centennial Flats phase three timeline has moved more than once.

The redeeming pattern, and it matters, is that bad news arrives early and unembellished. The impairment and target withdrawal were pre-announced two weeks ahead of the report. The Q2 2026 loss was attributed to its actual cause with no hedging. The Västvind write-down was disclosed in the first report after the rejection. The dividend suspension was explained mechanically, by naming the bond covenant, rather than framed as prudence.

The fair conclusion: a credible operator and an unreliable forecaster. Trust the transaction pipeline commentary; discount the multi-year targets, of which there are now fewer to discount.

10. Shareholder friendliness index

Dividends. Eolus paid SEK 2.25 per share for FY2023 (resolved at the AGM on 16 May 2024, record date 20 May 2024), SEK 2.25 per share for FY2024 in two installments of SEK 0.75 and SEK 1.50 (resolved at the AGM on 15 May 2025, record dates 19 May 2025 and 24 November 2025), and nothing for FY2025 (the AGM on 6 May 2026 resolved no dividend). For context, FY2021 and FY2022 each paid SEK 1.50, so FY2023 represented a 50% step up that was then held flat before being cut to zero. The cut is not discretionary in the ordinary sense: the company stated that "Eolus' negative result for the financial year 2025 prevents the payment of a dividend during 2026," because the terms of the senior secured green bonds (2025/2029, ISIN SE0024320774) contractually prohibit distributions following a negative annual result (Update regarding Eolus's capital allocation 2026). The stated policy remains 20 to 50% of group profit after tax, and it was explicitly reaffirmed on 27 January 2026 even as the operating profit target was withdrawn.

Buybacks and dilution. Over the full three-year window there has been no meaningful share repurchase. For the last ~90 days, per the recent record: the board resolved, under the authorisation granted at the AGM of 6 May 2026, to repurchase a maximum of 21,900 series B shares, with purchases able to commence on 29 August 2026 and run to the AGM 2027, administered by Carnegie, and explicitly for the purpose of hedging the Share Savings Program 2026 and its social security charges rather than for capital return (Eolus release). For the period older than 90 days, going back three years, web research of the AGM resolutions and company releases shows authorisations without execution: the AGM of 15 May 2025 authorised repurchase of up to 280,000 series B shares and transfer of up to 220,000 performance shares free of charge to employees, and the AGM of 6 May 2026 authorised repurchase of up to 10% of all shares. No general buyback programme was executed under either, and the company noted that repurchases would require bondholder approval to amend the bond terms. On dilution, the share count has been static: 1,283,325 class A and 23,623,675 class B, totalling 24,907,000 shares with share capital of SEK 24,907,000, unchanged since 2015 apart from small A-to-B conversions (1,000 shares in 2023, 1,300 in 2024) (Share Capital Development). The Share Savings Program 2026 permits a maximum of 125,000 performance shares, which would be 0.5% if fully delivered and is intended to be hedged by repurchase rather than issuance. One further note for completeness, since it is a capital return but not to shareholders: in May 2026 Eolus tendered for its own senior secured bonds up to SEK 200 million at 104.0% of nominal and repurchased SEK 71.25 million, leaving SEK 478.75 million outstanding.

Verdict: Returns Capital, currently blocked. Eleven years without issuing a share and a rising dividend through FY2024 make this a shareholder-friendly capital structure; the FY2025 zero is a bond covenant consequence of one impairment year rather than a change of intent.

11. Insider activities

Source: the Finansinspektionen PDMR transactions register (marknadssok.fi.se), the primary Swedish regulatory source under MAR Article 19, queried directly for Eolus over the twelve months to 19 September 2026. The register returned eight transactions in total for the period. Cross-checked against company releases and press reporting for anything more recent; no PDMR filings appear for Eolus between 15 June 2026 and 19 September 2026.

Recent transactions (most recent first)

All citations: Finansinspektionen PDMR transactions register, publication dates 20 Nov 2025, 24 Nov 2025, 16 Feb 2026, 12 Mar 2026, 1 Jun 2026, 3 Jun 2026 and 10 Jun 2026.

For context outside the twelve-month window, the CEO also bought 1,560 series B shares at SEK 45.20 on 27 August 2025, taking his holding to 78,866 shares at that time (MarketScreener, 27 August 2025). Per the company's group management disclosure, Witalisson holds 15,925 class A and 70,897 class B shares, and Christer Baden Hansen holds 17,296 class B shares (Group Management).

Buys - reading the signal

Six of the eight transactions in the window are share acquisitions, and five of them are by the CEO or the Deputy CEO buying in the open market with their own money. More important than the count is the timing, which is where the signal actually lives:

• 20 November 2025. The Q3 2025 report was published on 19 November and showed a loss-making quarter. The CEO and Deputy CEO both bought the following day. Two executives buying on the day after a bad print is cluster buying in the most meaningful sense: not coordinated scale, but coordinated conviction at the exact moment the market was being handed the bad news.
• 16 February 2026. The year-end report landed on 11 February, disclosing a loss-making year, approximately SEK 240 million of impairments, a withdrawn three-year profit target and no dividend. The CEO bought five days later.
• 1 June 2026. The CEO bought again, at SEK 44.025, having bought at SEK 36.93 in November and SEK 37.51 in February. He was buying into strength, not averaging down into weakness.

The CEO has bought in four separate windows in thirteen months (Aug 2025, Nov 2025, Feb 2026, Mar 2026, Jun 2026), on every occasion in the open market, and on two of those occasions within days of publishing materially bad news. This is a very bullish signal. The absolute sums are modest relative to a CEO's compensation, roughly SEK 300,000 across the twelve-month window, so this is not a bet-the-house purchase. But the pattern is what matters: a chief executive who has held the job since August 2012 and has been with the company since 2006, buying repeatedly and specifically after disclosing impairments, a withdrawn target and a suspended dividend, is making a statement that the write-downs do not change his view of the pipeline's worth.

The Deputy CEO's November 2025 purchase alongside the CEO's, on the same day and at a near-identical price, is the cluster element.

Sells - working out the why

Christer Hansen, 8 June 2026, disposal of 9,400 series B shares at SEK 42.08, followed on 10 June by an acquisition of exactly 9,400 series B shares at SEK 42.50. The identical volume, the two-day gap and the fact that he paid more to buy back than he received to sell make an outlook-driven sale implausible: he ended with the same position and slightly less cash. The most likely explanations are a transfer between custody arrangements, a move between a personal and an endowment or investment-savings account, or a related-party restructuring, all of which are reported as a disposal followed by an acquisition under MAR Article 19. Reason not disclosed in the register, and I am not going to assert one. What can be said with confidence is that it is not a reduction in exposure.

Hans Johansson, 3 June 2026, disposal of SEK 2,500,000 nominal of the Eolus Vind AB FRN05/29 bond at par. This is a debt instrument, not equity, so it says nothing about his view of the shares, and he remains the largest insider equity holder on the board with 189,520 class A and 47,111 class B shares per the company's board disclosure. The timing is the interesting part and points to a mechanical cause: Eolus ran a tender offer for exactly these bonds, announced 25 May 2026, up to a maximum nominal of SEK 200 million at 104.0% of nominal plus accrued interest, expiring 12:00 CEST on 29 May 2026, with settlement expected on or around 2 June 2026 and a total of SEK 71.25 million repurchased (Eolus release, 29 May 2026). A board member's bond disposal reported the day after that settlement date is most plausibly a tender into the company's own buyback. The register records the price as par, not the 104.0% tender price, so this inference is not confirmed by the filing itself and I flag it as inference rather than disclosure.

No insider sold a single Eolus share in the twelve months to 19 September 2026 in a way that reduced their equity position.

Net assessment

Insiders are unambiguously net buyers of equity, and there is no offsetting equity selling at all. The activity is concentrated in two people, the CEO and the Deputy CEO, which limits how broad-based a signal it is, but those are the two people closest to the project pipeline and to the impairment decisions. The single most telling feature is the timing: purchases on the day after a loss-making quarterly report, five days after a report disclosing SEK 240 million of impairments and a withdrawn three-year target, and again at a higher price four months later. The only disposals in the window are a same-size same-week round trip that left the position unchanged, and a bond sale that lines up with the company's own bond tender settlement.

One caution against over-reading: the sums are small in absolute terms, roughly SEK 300,000 for the CEO across the year, so this is conviction expressed cheaply rather than a founder-scale commitment. And the absence of any filings between mid-June and mid-September 2026 means there has been no insider vote on either the offshore rejection of 16 July or the Centennial Flats milestone of 14 August.

Read: bullish signal. Repeated open-market CEO buying immediately after self-disclosed bad news, with zero equity selling, is about as clean a signal as this section ever produces.

12. Scenarios

Bull case

The permitting winter ends, and the inventory Eolus has been quietly accumulating turns out to have been bought at the bottom.

The mechanism is unglamorous: Swedish municipal incentive payments rising to SEK 400 million by 2027 gradually shift the local political calculation, industrial electrification in northern Sweden finally catches up with the generation that was built for it, and Nordic forward prices lift off the floor. The 536 MW of fully permitted Swedish projects sitting without an investment decision start clearing, and behind them a pipeline of 5,834 MW of onshore wind becomes sellable again rather than merely permitted. Eolus, having entered the downturn with net cash, a 65% equity ratio and no need to dump assets, is one of the few developers that kept originating through the freeze. Its competitors did not: Arise was absorbed into Aneo, OX2 into EQT, and both spent the period integrating rather than prospecting.

Simultaneously, the storage business becomes the main event rather than the interesting sideshow. The Nordic energy storage market compounds at the projected low-teens rate towards USD 18 billion by 2035, the mFRR and ancillary service revenue stack proves durable, and the batteries Eolus is co-developing at the three former Mirova sites become the template: existing grid connection, existing land, existing operational relationship, short development cycle, quick sale. In the US, the asymmetry of the 2025 tax legislation works entirely in Eolus's favour, because storage retains its credits into the 2030s while everyone else's solar and wind pipelines hit the 2027 cliff. Eolus's four completed US storage divestments stop looking like opportunism and start looking like the core competence.

By 2028 or 2029 the company has done what it did in 2025 but in a market that pays more: several transactions a year, an asset management book pushing well past 1.5 GW with fifteen-year contracts underneath it, the bond covenant satisfied by consecutive profitable years, and the dividend restored and growing off a share count that has not increased since 2015. And the two peer takeouts have established what a Nordic development platform is worth to strategic capital, which means the option of being that platform is never entirely off the table.

Base case

Nothing is fixed, nothing breaks, and the company grinds through a bad cycle roughly the way it has been grinding through it.

Sweden stays hard. Municipal vetoes remain elevated, the 2027 build forecast of around 280 MW proves about right, and the Swedish onshore portfolio remains largely inventory. Offshore stays closed: the new tender regime takes years to design and run, defence screening remains the binding constraint, and Eolus does not re-enter it at scale, having already written down most of what it had there. The offshore line quietly disappears from the portfolio table.

The business Eolus actually transacts becomes the US, storage, and hybrids. Centennial Flats phase three closes out before the end of 2026 and delivers the last USD 12 million or so, Roccasecca reaches commercial operation, and the contingent USD 4.5 million on the September 2026 mid-stage sale partly arrives. Each year produces two to four transactions, most of them smaller than the Mirova or Roccasecca deals, enough to cover a fixed base of roughly 100 people and produce a profit in most years but not all of them. Quarters continue to swing violently between large positive and modestly negative, because that is what this model does, and the market continues to misread individual quarters.

Asset management keeps grinding upward, adding a few tens of MW a year off each divestment, and remains the only line anyone can forecast. The equity ratio stays far above the 30% floor because management keeps choosing liquidity over growth. A profitable year restores the dividend under the bond covenant, probably at the lower end of the 20 to 50% policy range, and the board keeps not buying back shares beyond the tiny amounts needed to hedge the employee share programme. Management sets no new multi-year profit target, having been burned by the last one, and investors judge the company deal by deal.

Bear case

The pipeline turns out to be a cost centre rather than an asset.

The sequence starts where it has already started. Swedish permitting does not improve, and the incentive payments prove irrelevant against local opposition that was never about money. Nordic prices stay near zero for longer than 2027, because electrification keeps disappointing and hydrology keeps cooperating. Fully permitted projects sit unbuilt not for one year but for four, and the accounting question becomes unavoidable: if nobody will take a final investment decision on a permitted Swedish wind project, what is the capitalised development cost of a permitted Swedish wind project actually worth? The Q4 2025 impairment of approximately SEK 240 million and the SEK 48 million Västvind write-down stop looking like isolated events and start looking like the first two entries in a series. Each one is a real loss, each one is announced in January, and each one resets the bond covenant clock so that the dividend never comes back.

The US does not compensate. Treasury's tightened begun-construction rules disqualify part of the solar pipeline from tax credits, the Foreign Entity of Concern regime makes a Swedish-owned developer a diligence problem for every tax credit buyer and lender, and the 2027 placed-in-service deadline passes with projects on the wrong side of it. Storage remains eligible but becomes the only thing Eolus can sell in the US, and it is competing for that narrow lane against Invenergy, Intersect, Arevon and the development arms of NextEra, all of which are larger and better positioned in interconnection queues. Sale prices per MW compress.

Meanwhile the fixed costs do not move. A development organisation of around 100 people across six countries cannot be scaled down without destroying the origination capability that is the whole asset, so management keeps paying for it out of a cash pile that the Mirova and Roccasecca transactions built and that nothing is replenishing. Two or three consecutive quarters like Q2 2026, with no closings and no progressive recognition, and the net cash position starts to look finite. The bonds at three-month STIBOR plus 750 basis points come due in 2029 into a credit market that has watched three years of impairments.

The final turn is the one that is hardest to see coming: the asset management book, the only stable earnings line, becomes a target. A larger operator or an original equipment manufacturer offers an institutional owner better terms across a bigger fleet, one fifteen-year mandate does not renew, and the stabiliser weakens at exactly the moment the development business cannot cover the overhead. At that point the strategic question is no longer how Eolus grows but who buys it, and the two peer takeouts establish that the answer is probably private capital, at a price set by a distressed seller rather than a confident one.