Skanska AB: A Record SEK 297.5bn Backlog Rests on a 4.3% Contractor Margin, and SEK 271.90 Leaves No Conservative Margin of Safety

The property side remains the main reason not to capitalize the new backlog aggressively. Commercial Property Development took SEK 668 million of write-downs in Q3 2025 and another SEK 464 million in Q2 2026 on a few completed unsold US properties. In Q2 management explicitly tied the new impairment to weaker market values, macro uncertainty and higher ten-year US Treasury rates. Two impairment rounds separated by three quarters are economically a running re-marking process, even if the individual accounting charges concern discrete buildings. I would resist calling them isolated one-offs. Skanska still carried SEK 36.18 billion of Commercial Property Development assets at June 30, against separately disclosed components of SEK 18.45 billion of completed projects, SEK 6.50 billion of investment in ongoing projects and SEK 11.31 billion of land and development properties.

There is an important counterweight. Skanska estimates the CPD portfolio's market value at completion at SEK 46.74 billion against investment value upon completion of SEK 42.58 billion, implying SEK 4.16 billion of unrealized gains. Residential Development likewise had SEK 3.4 billion of estimated pre-tax surplus value in unsold homes under construction and land. Those figures argue against treating the property portfolio as impaired capital in its entirety. Yet they are management estimates whose reliability depends on yields, leasing and transaction liquidity. After repeated US marks, I would value CPD below carrying value in a base case and require actual divestments to close the gap before paying for the reported surplus.

Nordic residential development is in a different phase. Skanska had 2,878 homes under construction at June 30, 55% sold, while completed unsold homes had fallen to 326 from 516 a year earlier. That inventory improvement is encouraging. New activity remains subdued: Q2 starts were only 225 homes versus 420 a year earlier, and management described Norway and Finland as weak while Sweden was improving. Sweden's Riksbank kept its policy rate at 1.75% in August 2026, and the average floating rate on new Swedish household mortgages was down to 2.74% in July. Swedish housing starts rose sequentially in Q2, but activity remains far below the levels of the previous housing cycle. Lower financing costs have removed part of the pressure; they have not restored old volumes.

Capital allocation looks more disciplined than the headline SEK 14 dividend suggests. The ordinary component was SEK 8.50 a share, equivalent to 56% of 2025 EPS and comfortably within Skanska's policy of distributing 40–70% of net profit when its financial position is stable. The SEK 5.50 extra lifted the total payout to 93% of 2025 EPS. I do not regard that extra as recurring income. It was proposed on February 6 and approved March 31; the approximately USD 75 million I-4 Ultimate stake divestment was announced/completed in early April and therefore was not the specific funding event behind the February dividend proposal. The more plausible source was excess balance-sheet capacity after years of cash retention. The Q2 balance sheet remained in adjusted net cash of SEK 8.7 billion even after the dividend.

Skanska's qualitative portrait is best described as a company in transition from a cyclical mixed-quality contractor-development conglomerate toward a more selective contractor carrying a still-cyclical property balance sheet. The Construction repair has progressed far enough to be visible in four years of margin data; Project Development has not yet earned its 10% return-on-capital target, with the Q2 rolling return at only 1.3% for Project Development and 4.8% for Investment Properties against a 6% target. The market's debate is unusually clean: bulls see the 4.3% margin as a floor beneath an unprecedented technology/infrastructure book; bears see current Construction profitability as close to peak while SEK 50-plus billion of development carrying value remains exposed to rates and transaction markets.

At SEK 271.90, the Class B share is materially above its SEK 252.30 2025 year-end level and has recovered more than 10% from the Q2-result close. The price is also above the recent 16–17 times trailing-EPS valuation center: at rolling EPS of SEK 15.69, it trades at about 17.3 times earnings and about 1.64 times Q2 adjusted equity per share of SEK 166.25. Those ratios are useful context but poor valuation tools for the whole company. The decision turns on what Construction is worth on normalized earnings plus how heavily the development balance sheet should be discounted.

Vertical history, financial record and market narrative

Skanska began in 1887 as Skånska Cementgjuteriet in southern Sweden, initially commercializing concrete construction at a time when industrialization and urban infrastructure were expanding. The business gradually moved from producing and applying concrete into general contracting. That origin matters because Skanska's historical advantage has never rested on a proprietary product in the modern technology-company sense. Its recurring asset has been the capability to price, organize, finance and execute increasingly complex physical projects.

The first stage ran from the late nineteenth century through the 1960s: domestic scale and engineering breadth. By 1964 sales exceeded SEK 1 billion, and Skanska listed on the Stockholm Stock Exchange's A-list in 1965. The company's own historical material confirms the listing but does not disclose enough contemporaneous primary-source detail to reconstruct a modern-style IPO price, proceeds or book-building valuation. I do not invent one. The capital-market story was a mature industrial builder entering the public market, rather than a venture-funded company monetizing a new category.

The second stage was internationalization. Skanska participated in Sweden's large post-war housing build-out and, in 1971, entered the United States with projects including subway work in New York and Washington. The company was renamed Skanska in 1984. Over the following decades it became a multinational builder and developer, with the US eventually becoming its largest single market. Major bridges, transport systems and complex civil projects created the credentials that still underpin today's ability to bid for multibillion-krona US infrastructure work.

International scale also created the company's recurring weakness: construction revenue can grow faster than the organization's ability to price project risk. Skanska's own history includes the Hallandsås tunnel environmental crisis in the 1990s, while later US public-private-partnership and civil projects generated substantial losses. The 2010s culminated in a reassessment of bidding discipline and portfolio exposure. Anders Danielsson, who had substantial operating experience inside Skanska including in the United States, became group CEO in 2018. The strategic legacy is visible today in management's emphasis on project selection, cash-positive profiles and a minimum Construction margin rather than revenue maximization.

The third stage, roughly 2018 through the pandemic period, was a risk reset. Skanska reduced its appetite for projects whose contract terms offered inadequate compensation for design, execution or joint-venture risk. This changed the economic objective from “fill the order book” to “fill it with work that protects cash and margin.” The evidence is not a single clean upward line, because the pandemic and project-development sales distorted group earnings, but the current backlog is being built on top of a Construction profitability base that is stronger than the one Skanska carried into the late 2010s. The free-working-capital discipline management stresses today is part of that same reset.

The fourth stage began with the 2022–2023 rate shock. It hit the two halves of Skanska asymmetrically. Construction benefited from public infrastructure, healthcare, industrial and eventually data-center demand while development activity lost transaction liquidity and faced higher required property yields. The income statement captured this divergence in 2023: Construction still achieved a 3.5% operating margin, but Project Development swung to a roughly SEK 2.6 billion operating loss. Residential volumes fell and commercial development became harder to sell. The stock had to be valued less as a broad construction boom beneficiary and more as a contractor whose development assets were temporarily consuming capital and confidence.

From 2024 onward, Construction became the repair engine. The margin held at 3.5% in 2024, rose to 4.1% in 2025 and reached 4.3% on a rolling basis by June 2026. Project Development returned to profit in 2024 and 2025 but at returns far below its capital target, and repeated US commercial-property impairments prevented the recovery from becoming a clean normalization story. This is the current stage: the contracting business has repaired its economics faster than the balance-sheet development businesses have repaired theirs.

The table captures why a group P/E is unstable. EPS halved in 2023 even though Construction's margin held at 3.5%, because development overwhelmed the contractor's earnings. The reverse can also occur: a property-sale year can make group earnings look unusually cheap while the recurring construction economics have barely changed.

Vertically, the cash-flow record sends the same warning. Q2 2026 operating cash flow of SEK 7.13 billion exceeded the quarter's SEK 1.79 billion profit by four times, but SEK 6.15 billion of the reported cash generation came from the combination of working-capital release and net divestments. On a rolling twelve-month basis OCF was SEK 8.34 billion against SEK 6.57 billion profit, a healthy 1.27 times. In 2025, the equivalent ratio was only about 0.57 times. This volatility is structural because development investment and sales, customer advances and contract-working-capital movements sit alongside operating profit.

A clean five-year OCF/net-income ratio cannot be reproduced to investment-grade precision from the primary extracts available in this research session without mixing Skanska's segment “operating cash flow from operations” definition with IFRS cash-flow classifications for older periods. The accessible series nevertheless shows the essential economic point. Conversion is highly uneven across years, with weak 2023 cash generation, strong 2024, a softer 2025 and a strong R12 through June 2026. I therefore do not value the group on a single-year FCF yield; Construction is valued on normalized owner earnings while development assets are valued on carrying/NAV measures.

Construction capex is much more tractable. Over the rolling twelve months, gross Construction investments were SEK 3.37 billion and divestments SEK 0.32 billion, for net investment of SEK 3.05 billion. Q2 depreciation of PP&E and right-of-use assets was SEK 655 million, an annualized run rate near SEK 2.6 billion. Skanska does not disclose a precise maintenance-versus-growth split. I estimate roughly SEK 2.5–2.8 billion as maintenance-like spending and the remainder as growth/renewal; the range is deliberately close to depreciation because Construction's asset base is mature. Under that assumption normalized Construction owner earnings are close to after-tax operating earnings rather than materially below them.

The development balance sheet grew even while profitability remained weak. At June 30, Residential Development had SEK 17.77 billion of carrying amounts, including SEK 10.91 billion of undeveloped land and development property. CPD had SEK 36.18 billion of carrying amount, while Investment Properties were valued at SEK 8.29 billion. Capital employed is a different measure from those carrying amounts and is not their sum: across RD, CPD and IP it totaled SEK 63.38 billion. Against that, the group held adjusted interest-bearing net receivables of SEK 8.7 billion and total equity of SEK 61.7 billion. This is why “net cash” should not be confused with “capital light”: the group is liquid, but much of shareholder capital is still committed to property.

Governance has been stable. The current Group Leadership Team comprises CEO Anders Danielsson; CFO Pontus Winqvist; and executive vice presidents Lena Hök, Richard Kennedy, Claes Larsson, Ståle Rød, Therese Tegner and Åsa Thunman. No primary-source announcement of a CEO succession was located as of the research date. The board remained chaired by Hans Biörck after the March 2026 AGM.

Ownership is unusually stable for a large contractor. Industrivärden and Lundberg are major long-term holders, and the Class A voting structure gives such holders greater voting influence than their economic stakes alone would imply. At June 30, Industrivärden controlled about 24.9% of votes with 8.1% of capital and Lundberg about 13.3% of votes with 5.9% of capital. That structure can reduce takeover optionality and gives established Swedish owners substantial governance influence; conversely, it provides an owner base capable of supporting multi-cycle project discipline rather than quarterly revenue maximization.

Capital allocation in 2026 deserves a precise reading. The ordinary SEK 8.50 dividend alone represented about 56% of 2025 EPS, squarely inside the 40–70% policy. The extra SEK 5.50 brought the total distribution to SEK 14 and 93% of EPS. Skanska has paid extra dividends before, including in 2020 and 2021, so such payments are not unprecedented. They are not a recurring entitlement. I would model SEK 8.50 as the sustainable starting distribution and give no base-case value to another SEK 5.50 extra.

The I-4 Ultimate transaction reinforces that view. Skanska sold its 50% stake for approximately USD 75 million in April. Using the company's own rounded conversion at announcement, that was about SEK 690 million, an implied USD/SEK rate of roughly 9.2 on the transaction disclosure date. The gain appeared in Central in Q2. It is too small to explain the roughly SEK 2.3 billion cash cost of the extraordinary dividend, and the dividend proposal preceded completion of the sale. It was a favorable monetization of mature P3 capital, not the economic source of a permanently higher payout.

The Class B price history over the latest cycle captures the shifting narrative. Skanska ended 2024 at SEK 232.70, traded between SEK 185.40 and SEK 268.10 during 2025, and finished 2025 at SEK 252.30. The 2026 Q2 results pushed the stock down to roughly SEK 246.70 on July 17; it was around SEK 262.30 before the largest August data-center award and closed September 4 at SEK 271.90. The last leg therefore represents a re-rating after the earnings disappointment, helped by repeated order announcements and a broader acceptance that Construction's book is unusually strong.

The Q3 2025 CPD impairment did not create a durable collapse in the shares: Skanska still ended 2025 at SEK 252.30. That suggests investors separated construction improvement from property marks. The March/April 2026 dividend period also needs total-return adjustment: the share went ex a SEK 14 total dividend on April 1, and Nasdaq explicitly adjusted derivatives for the SEK 5.50 extraordinary component. A mechanical ex-dividend decline is not evidence that the investment thesis worsened that day.

The longer valuation history is cyclical. Year-end P/E was about 16.5 times in 2024 and 16.7 times in 2025. It exceeded 23 times in 2023 because development losses depressed the denominator, while 2021–2022 year-end P/Es were much lower because EPS was elevated relative to the subsequent cycle. At SEK 271.90 against rolling segment EPS of SEK 15.69, today's 17.3 times is modestly above the latest normalized range, not an obvious trough multiple. That is another reason the stock requires asset-by-asset valuation rather than a “low versus history” claim.

Business model, moat and industry cycle

The business machine starts with Construction. Customers award projects to Skanska, often years before completion. Skanska recognizes revenue as work progresses, pays subcontractors, labor and suppliers, and seeks to structure projects so customer advances and milestone payments precede cash outlays. The economic spread is narrow. In the R12 period Construction had an 8.3% gross margin, 4.0% SG&A ratio and 4.3% operating margin. Small errors in tender price, schedule or subcontractor cost can erase a large part of project profit.

Project Development runs the opposite cash cycle. Skanska buys or controls land, designs and permits projects, invests in construction and leasing, and then realizes profit when homes or properties are sold. Accounting earnings arrive in lumps because asset transfers are discrete. CPD illustrates the model: R12 revenue was SEK 9.53 billion, but the balance sheet carried SEK 36.18 billion of current asset property. The metrics that matter are return on capital, leasing, occupancy, value relative to cost and the ability to recycle capital, rather than revenue growth by itself.

Segment revenues include internal items and therefore do not add directly to consolidated group revenue. The profitability split is more revealing than the revenue split: Construction currently produces nearly all recurring operating profit, while more than SEK 60 billion is tied to businesses earning far below Skanska's stated return targets.

The Construction cost base is largely variable at the project level because subcontractors, materials and project labor scale with activity. Corporate engineering, bidding teams, management systems and local branch infrastructure create a smaller fixed layer. Scale does not create software-like operating leverage. More volume helps only when the added contracts are at least as well priced as the existing book. A contractor that chases volume can increase revenue and destroy margin simultaneously. Skanska's 2018-era risk reset and the subsequent improvement to a 4.3% margin suggest management has learned that lesson.

The strongest moat is qualification and execution credibility on large, difficult projects. A public transport authority, hospital network or hyperscale technology client does not choose a general contractor solely on nominal bid price; bonding capacity, safety record, relevant references, schedule confidence, subcontractor networks and the ability to absorb a multi-year project matter. Skanska's Q2 awards included a SEK 9.3 billion Massachusetts bridge project, SEK 4.6 billion New York subway work and SEK 4.3 billion Gateway tunnel package. Such awards show access to a market whose practical barriers are materially higher than ordinary commercial building.

A second real advantage is repeat-customer trust in technology buildings. Skanska markets a history of more than 250 data-center projects and substantial mission-critical capability. More importantly, several 2026 data-center releases explicitly call the buyer an “existing client.” Repeat business does not reveal pricing power, but it reduces customer-acquisition friction and shows that delivery history is passing the customer's technical and procurement tests.

A third advantage is the balance sheet. A net-cash contractor can bid large projects, post guarantees, withstand disputes and fund temporary working-capital swings without being forced into expensive equity or distressed borrowing. Skanska had SEK 23.1 billion of cash, short-term investments and committed unused facilities at June 30 and adjusted net receivables of SEK 8.7 billion. That financial capacity matters more in construction than headline leverage ratios imply because joint-venture commitments and project guarantees can become cash claims when projects go wrong.

The moat is medium rather than strong because customers still control price and contract terms. Skanska has no patent, network effect or proprietary technology that forces a customer to buy from it. Its edge is a combination of reference projects, risk processes, balance-sheet capacity and people. Those assets can sustain a few hundred basis points of good execution, but they do not remove project-loss risk. A single mispriced mega-project can consume years of ordinary margin.

Data centers sharpen that distinction. The end market has characteristics Skanska likes: very large projects, demanding mechanical/electrical interfaces, schedule sensitivity and customers who value repeat execution. Yet specialized peers illustrate where the profit pool can sit. Sterling Infrastructure's mission-critical E-Infrastructure business has reported adjusted operating margins above 20%, far above Skanska's group Construction margin, because site development and electrical scope are more specialized and carry different pricing economics. Skanska is participating in the same capital-expenditure wave through a general-contractor model, but that does not make its economics equivalent.

The data-center order wave also sits inside a diversified backlog. Q2's five largest disclosed orders included bridges, subway work, a hospital, a tunnel and a commercial office. That diversification matters because AI-related capital expenditure can be volatile. Public infrastructure generally responds to government budgets and long planning cycles; semiconductor and data-center projects respond more directly to private technology capex and power availability. Combining the two produces a healthier order book than pure hyperscale exposure, provided one category does not become dominant.

The exact data-center concentration cannot currently be measured from Skanska's disclosure. Four data-center contracts disclosed for Q3 through August 26 totaled SEK 15.16 billion: SEK 11.2 billion for four facilities in the southeastern US, SEK 2.2 billion for a Virginia campus addition, SEK 930 million near Prague and SEK 830 million in Finland. The US awards describe anonymous existing clients, Prague identifies CRA Prague Gateway DC, and Finland identifies only a technology company. Public disclosure does not establish how many ultimate hyperscalers sit behind the anonymous awards.

Two contracts in the original research brief require reclassification. The SEK 870 million Virginia contract announced July 7 and the SEK 2.3 billion Georgia award announced June 23 are explicitly listed by Skanska as Q2 bookings. A second SEK 870 million Virginia data-center award announced July 2 is also listed in Q2. Announcement date and accounting booking quarter are not interchangeable.

The disclosed Q3 data-center awards alone equal 5.1% of the June 30 backlog. Add Q2's SEK 6.7 billion of data centers and semiconductor facilities and the recent technology-related booking flow reaches SEK 21.86 billion, equivalent to 7.4% of Q2 backlog. That 7.4% is not the actual technology share of backlog: some Q2 work will already have converted to revenue, semiconductor work is mixed with data centers, and older technology projects remain in the book. The correct conclusion is that data centers are already material but Skanska does not publish the stock measure needed to quantify the category precisely.

A useful concentration stress test is the two large anonymous US Q3 awards. Together they represent SEK 13.4 billion, or 4.5% of June backlog. If they belong to the same hyperscaler and that client pauses future phases, the visible growth rate of backlog could lose several percentage points even before considering earlier anonymous projects. Signed backlog would not necessarily disappear dollar-for-dollar because termination compensation, work already performed and contract clauses matter. Those clauses have not been disclosed. This is a concentration risk that cannot be quantified away with the word “record.”

Contract form is the largest missing variable in the AI thesis. The award releases disclose value, location, customer descriptors and completion timing but not whether the projects are lump-sum/fixed-price, guaranteed-maximum-price, construction-management-at-risk or cost-reimbursable. Skanska also does not disclose the overall backlog percentage by contract form. Without that information, it is impossible to prove that a surge in technology awards has reduced or increased cost risk. I therefore give zero valuation credit to a hypothetical data-center margin premium until segment/geographic margin data reveal it.

The Nordic housing cycle is different. Swedish monetary policy has moved from restraint toward neutrality: the policy rate was 1.75% as of late August, and household mortgage rates have fallen. Skanska's Swedish sales are improving, but Norway and Finland remain subdued and Q2 Residential operating margin fell to 1.5% from 11.3% a year earlier. The decline partly reflected low volumes, weak-margin legacy projects, warranty provisions and SEK 70 million of restructuring/warranty costs. Lower rates create a recovery option; they do not erase the low-return land capital already on the balance sheet.

Commercial property responds to the opposite end of the rate curve. Central-bank cuts can improve residential affordability while long government-bond yields remain high enough to depress office capitalization values. That is exactly what Skanska cited in Q2 for its US CPD impairments. Leasing can therefore look acceptable while sales prices remain disappointing. At June 30 the CPD portfolio's completed projects had an estimated market value only SEK 822 million above SEK 18.45 billion of investment value. That thin cushion is why completed US assets deserve closer scrutiny than the broader portfolio's SEK 4.16 billion reported surplus implies.

Investment Properties is smaller and cleaner: SEK 8.29 billion of property value, 83% economic occupancy, a 4.7% average valuation yield and a 73% surplus ratio, meaning net operating income as a share of rental income rather than unrealized gain. Management aims eventually to build the Swedish office portfolio to SEK 12–18 billion. I would prefer restraint here until Project Development returns improve. Expanding a balance-sheet property portfolio while CPD and RD are earning far below targets would consume capital that the Construction business does not need in order to grow.

Horizontal competitor analysis

The most useful horizontal comparison separates three species rather than pretending every construction ticker is interchangeable. Skanska and Peab are Nordic builders with meaningful development/property exposure. ACS/Hochtief through Turner shows what a large global construction-management platform can become when digital infrastructure dominates growth. Sterling, MasTec and MYR show how much higher margins can be when the contractor owns specialized site, electrical, utility or network scopes. Ferrovial sits outside that continuum because infrastructure concessions and ownership economics dominate its investment case; it is useful mainly as a reminder that “infrastructure” does not imply contractor margins.

ACS has become increasingly North American and technology-infrastructure driven through Hochtief and Turner. Turner reported first-quarter 2026 new orders up 48%, work-in-place revenue up 25% and backlog up 34%, with data centers a principal driver. ACS separately reported Turner's EBITDA margin rising 72 basis points to 3.9%. That last figure is instructive: even an exceptional data-center construction-management franchise still operates on single-digit contractor economics. Turner's data-center wave has coincided with margin expansion, but its disclosure proves Turner's experience, not Skanska's.

Hochtief's first-half backlog reached a record EUR 84.8 billion, up 23%, with new orders equivalent to 1.5 times work completed. Digital infrastructure orders more than doubled over twelve months. In other words, Skanska's 114% book-to-build and record backlog are not occurring in isolation; there is a genuine global wave of mega-project awards. The comparison weakens the argument that Skanska alone has discovered a new structural moat. It strengthens the argument that capable general contractors are experiencing a rare demand environment.

Sterling Infrastructure is the better contrast for margin quality. Its Q2 signed backlog was up 116% year on year and combined backlog up 150%. In Q1, more than 90% of E-Infrastructure backlog came from mission-critical projects including data centers, manufacturing and semiconductors. Its adjusted E-Infrastructure operating margins are around the mid-20s, many times Skanska's Construction margin. Customers choose Sterling for specialized site and increasingly electrical scope rather than for a complete general-contractor offering. The higher margin is therefore economically earned through a narrower, more specialized piece of the project.

MYR Group occupies another high-skill layer: transmission and distribution plus commercial and industrial electrical work. MYR reports in dollars, so at June 2026 the comparable measure is growth rather than a SEK amount: backlog rose about 20% year on year; first-half operating margins were 9.6% in T&D and 8.3% in C&I. MYR can benefit from the same power-intensive data-center build-out even when it is not the prime general contractor. Its economics support a key industry conclusion: AI infrastructure's richer profit pool often sits with constrained electrical and power capabilities, not automatically with whoever records the largest construction contract value.

MasTec provides a broader US infrastructure comparison. Q2 2026 revenue grew 23% to a record level and its 18-month backlog reached a record USD 21.4 billion, driven especially by Clean Energy and Infrastructure. The model is more specialized than Skanska's and has substantial power, communications and energy exposure. For Skanska, this peer confirms that infrastructure capex is broadening beyond data centers themselves into the networks needed to serve them. It does not justify importing MasTec's margins or multiple into a Nordic contractor-development SOTP.

Peab is the cleaner Nordic operating comparison. Q2 2026 sales rose 12%, operating margin reached 5.5%, orders were SEK 18.34 billion and backlog a record SEK 57.29 billion. The first-half margin was only 2.8%, illustrating Nordic construction seasonality. Peab's development and property exposure also means investors must look beyond contractor margin. Skanska's 4.3% rolling Construction margin is less seasonally flattering than Peab's single-quarter 5.5%, and Skanska's US scale gives it access to a technology/infrastructure growth pool Peab largely lacks.

† Sterling's >90% figure includes data centers, manufacturing and semiconductor facilities rather than data centers alone.

The horizontal lesson is that Skanska's 4.3% is a good general-contractor margin, not an extraordinary infrastructure margin. Turner shows that technology mix can improve a construction manager by tens of basis points; Sterling and MYR show what happens when a company controls specialized scopes where labor, engineering and capacity are tighter. Skanska's plausible margin upside should be measured in tenths of a percentage point, not imagined as a path toward specialty-contractor margins.

Skanska's niche is “large-project integrator with own-development optionality.” It competes for complex US and European projects where scale, execution record and balance sheet matter, but it also invests its own capital in property. That differentiates it from Turner, whose construction-management exposure is cleaner, and from Sterling/MYR, whose specialty economics are richer. It is also why Skanska deserves a conglomerate-style SOTP discount whenever development returns are poor: the Construction franchise can be operating well while property capital destroys the group's return on equity.

The comparison also clarifies why Ferrovial is a poor direct multiple reference. A concession owner earns long-duration cash flows from infrastructure assets and accepts financing/regulatory risk; Skanska mainly earns construction margins and property-development gains. A higher concession multiple does not establish undervaluation at Skanska. Saint-Gobain is farther upstream: materials pricing, industrial utilization and distribution economics dominate. Both are useful for understanding construction demand, neither should determine the Class B fair value.

Exact base-date peer P/E and EV/EBITDA comparisons are intentionally not used as the primary valuation anchor here. The operating models are too different, and this research session did not retrieve a consistent September 4 primary-market valuation set for every peer. The safe inference from fundamentals is narrower: Skanska's Construction margin is already around the upper end of its own recent history but below specialty-contractor economics; Turner's own 3.9% EBITDA margin prevents treating 4.3% as obviously low. The absolute valuation below stands on its own.

Current fundamentals, valuation, risks and tracking

Q2 2026 was fundamentally stronger than the headline revenue reaction suggested. Group revenue was SEK 44.56 billion versus SEK 44.55 billion a year earlier and grew 1% in local currencies. Operating income increased 17% to SEK 2.12 billion, profit to SEK 1.79 billion and segment EPS to SEK 4.27 from SEK 3.69. Construction generated SEK 1.83 billion of operating profit at a 4.3% margin. Residential Development earned only SEK 25 million; CPD lost SEK 170 million after the SEK 464 million impairment; Investment Properties earned SEK 85 million; Central contributed SEK 334 million, helped by I-4.

The revenue miss mattered because the market had already begun paying for backlog conversion. Currency obscured some of the underlying trend: R12 group revenue was down 3% in SEK but up 3% in local currency, a six-percentage-point translation effect; R12 operating income rose 10% in SEK and 15% locally. A strong krona can make reported growth look materially weaker than operating growth without changing local project profitability.

Q2 cash flow is best reconstructed line by line. SEK 1.61 billion came from business operations, SEK 2.23 billion from working-capital improvement and SEK 3.92 billion from net divestments. After taxes and financing, OCF was SEK 7.13 billion. Four previously sold CPD projects were handed over, while Construction received milestone and mobilization payments. This cash is real; its repeatability is lower than the headline.

The working-capital position is one of Skanska's underappreciated assets. Free working capital in Construction reached SEK 33.9 billion and averaged 18.9% of Construction revenue over the preceding twelve months. Contract liabilities on the IFRS balance sheet were SEK 31.48 billion against SEK 12.13 billion of contract assets. The contractor is, in aggregate, being financed in part by customers rather than financing them. That arrangement increases return on capital when execution is good and creates cash risk when production or payment profiles reverse.

Backlog quality, rather than backlog size, is the next earnings question. The book has 21 months of production visibility and contains large public infrastructure projects plus a rapidly growing technology component. What remains undisclosed is the contract-risk composition, end-client concentration inside anonymous data-center awards and the expected margin of those projects. The next few quarters need to show revenue acceleration without Construction margin falling below 4% before the market can treat the record book as a durable earnings step-up.

The property recovery is behind schedule. Project Development's rolling ROCE was 1.3% in Q2 versus management's at-least-10% target, while Investment Properties generated 4.8% against a 6% target. Residential's completed unsold inventory has improved, but the business still carries SEK 10.9 billion of undeveloped land and development property. CPD's operating result remains vulnerable to appraisal and transaction-market shifts even when leasing progresses.

Repeated US write-downs deserve a direct answer. Q3 2025 CPD recorded SEK 668 million of write-downs/reversals; Q2 2026 added SEK 464 million on completed unsold US properties. The cumulative two-round charge is approximately SEK 1.13 billion. I regard the accounting entries as asset-specific but the economic driver as recurring: long rates and transaction values have repeatedly forced Skanska to lower carrying expectations. A third meaningful US charge would confirm that the marks are a continuing portfolio re-pricing rather than cleanup.

What remains cannot be answered at the granularity the ideal investor would want. Skanska reports SEK 18.45 billion of completed CPD project investment value, but the Q2 table does not separate the US subset of completed unsold assets. Total CPD carrying value is SEK 36.18 billion. The company estimates the completed portfolio's market value at SEK 19.27 billion, only SEK 0.82 billion above investment value; the larger reported surplus sits mainly in projects under development. I therefore apply a substantial discount to the entire CPD carrying amount in the valuation instead of assuming the US marks are finished.

The balance sheet makes that caution affordable. Adjusted net receivables were SEK 8.7 billion at June 30, versus SEK 9.5 billion at March 31, while total liquidity plus committed unused facilities was SEK 23.1 billion. Central borrowings were SEK 7.6 billion, with an average maturity of 1.3 years for the MTN portion and 2.5 years for bilateral loans. Skanska is not facing a refinancing problem; the question is the return earned on the capital already committed to development.

Valuation begins with cash passthrough. Rolling operating cash flow of SEK 8.34 billion divided by rolling profit of SEK 6.57 billion gives 1.27 times cash conversion. The same ratio was 0.57 times in 2025. The Q2 quarter itself was about 4.0 times, almost entirely because of working capital and property handovers. That dispersion makes reported FCF yield a poor single metric.

For owner earnings I isolate Construction. R12 Construction EBIT was SEK 7.18 billion. Applying an approximately 20–21% normalized tax burden produces about SEK 5.7 billion after tax. Construction's R12 net investments of SEK 3.05 billion were only modestly above an annualized depreciation run rate in the mid-SEK 2 billions, suggesting maintenance capital is broadly covered by depreciation. That puts my normalized construction owner-earnings estimate at roughly SEK 5.6–5.9 billion. The corresponding earnings multiple becomes meaningful only after the development assets and net cash are stripped out through SOTP.

Headline group numbers are 17.3 times rolling segment EPS, a 5.8% accounting earnings yield, versus about 1.64 times adjusted equity. The ordinary SEK 8.50 dividend implies a 3.1% yield at the current price; using the SEK 14 paid in 2026 gives 5.1%, but that would incorrectly capitalize an extraordinary distribution as recurring.

The SOTP uses Construction owner earnings, not consolidated net income. Development takes explicit discounts to carrying value. CPD absorbs the largest haircut, because two US impairment rounds have reduced confidence in the remaining balance-sheet marks. Residential gets a smaller one: completed inventory is declining and the portfolio includes SEK 3.4 billion of management-estimated pre-tax surplus. Investment Properties sits near book because it is already fair-valued at a disclosed 4.7% yield, though occupancy is only 83%. Net cash is added at par; normalized Central cost is capitalized as a deduction.

This is valuation-scenario analysis within a research framework, not investment advice.

The conservative construction case assumes the recent improvement slips below management's 4% target and applies only 10.5 times owner earnings. CPD is marked to 65% of carrying value, which embeds materially more pain than management's appraisal tables currently show. That yields roughly SEK 237 per share. It is a downside valuation, not a distress liquidation estimate.

The base case keeps Construction near the current 4.3% margin. That is deliberately different from assuming data centers create another margin leg. RD is valued at 90% of carrying amount, CPD at 82% and Investment Properties at 95%. The resulting roughly SEK 313 per share says the current price already discounts some development recovery but does not fully price a strong realization of stated property values.

The optimistic case requires the data-center/infrastructure mix plus project discipline to push Construction toward 4.7%, while property marks prove conservative and transaction liquidity returns. Even then I do not value CPD above carrying value despite Skanska's reported unrealized gains. At roughly SEK 391 per share, the bull case depends on both sides of the group working simultaneously.

The most fragile base-case assumption is normalized Construction owner earnings. Cutting that assumption to 70% while leaving the rest of the SOTP unchanged reduces base value from roughly SEK 313 to about SEK 261 a share. That sensitivity is important: a 4.3% margin sounds low in absolute terms, but the business generates such enormous revenue that one percentage point of margin is worth roughly SEK 1.7 billion of annual operating income. Project selection remains the dominant valuation variable.

The margin-of-safety test comes out unfavorable at the current quote. SEK 271.90 is about 15% above the conservative SOTP of roughly SEK 237. There is no discount to the conservative value, so by definition there is no conservative-case margin of safety. A three-year flat-earnings case in which the valuation multiple stays around today's level and the ordinary SEK 8.50 dividend is repeated produces only roughly 3% annualized total return before reinvestment assumptions. That return is too small an equity premium for a business retaining material project and property-cycle risk.

Margin-of-safety sufficiency verdict: none.

The price is nevertheless below base intrinsic value, which makes the distinction between “no margin of safety” and “overvalued” important. The current quote can generate a reasonable return if Construction remains at or above 4% and development returns recover. An investor is simply being paid too little to assume the conservative case is wrong. That leads to a hold-zone rather than a bargain-zone classification.

The main business risk has medium probability and high impact: Construction margin reversion. The observable signal is rolling margin falling below 4% or a large geography reporting project write-downs. The transmission path is immediate because every 50 basis points of margin on roughly SEK 166–170 billion of revenue represents about SEK 0.8–0.9 billion of EBIT. A move back toward 3% would erase much of the valuation premium earned during the recent risk-discipline period.

The second risk has medium probability and high impact: hidden data-center customer concentration. The visible US Q3 awards alone total SEK 13.4 billion and could conceivably rest on one anonymous existing client; public disclosure does not allow that to be resolved. A hyperscaler pause would first hit new bookings and production scheduling, then potentially working capital and Construction revenue. The stock's AI/order narrative would probably compress before accounting earnings moved materially.

The third risk has medium-to-high probability and medium-to-high impact: further US commercial-property markdowns. Two rounds totaling roughly SEK 1.13 billion have already occurred. A further quarterly impairment above roughly SEK 0.5 billion, especially on completed projects, would challenge the credibility of the remaining SEK 36.18 billion CPD carrying value and force a larger SOTP discount.

The fourth risk is a working-capital reversal. Probability is medium; impact can be high on cash and sentiment even when accounting earnings initially hold. Free Construction working capital of SEK 33.9 billion is an asset only while milestone timing remains favorable. A decline in average free working capital below roughly 15% of Construction revenue, accompanied by negative operating cash flow, would indicate that customers are financing less of the production cycle.

The fifth risk is a prolonged Nordic residential slump. Its group impact is lower than a Construction failure but it can trap capital for years. Completed unsold inventory falling from 516 to 326 is moving the right way, yet starts remain low and Norway/Finland weak. A renewed rise above 450 completed unsold homes, combined with a sales rate below 50%, would indicate that falling mortgage rates have failed to restart demand.

FX is visible but less likely to cause permanent loss by itself. R12 translation reduced reported revenue growth by six percentage points relative to local-currency growth and operating-income growth by five points. A stronger krona suppresses reported USD earnings and backlog; it does not necessarily weaken US project economics. Investors should therefore judge Construction order and profit growth in local currency first, then translate to SEK for valuation.

Positive catalysts are concrete. Q3 has already accumulated a large order slate, with the SEK 11.2 billion southeastern US data-center package and SEK 8.9 billion LA Metro award among the largest. The stronger catalyst would be Q3 revenue growth accelerating while Construction margin remains 4.3% or better. A CPD quarter with asset sales and no further impairment would help the property discount. Continued decline in Nordic completed unsold homes would make the RD surplus more credible.

Negative catalysts are equally measurable: cancellation or delay of a major technology award, Construction margin falling below 4%, another US CPD charge around SEK 0.5 billion or larger, and a sharp working-capital outflow after the current mobilization-payment benefit. The expected next earnings report is the Q3 2026 interim report on November 5, 2026.

The dashboard should be read as a causal system. Construction margin and working capital show whether backlog quality is translating into economic value. CPD impairments show whether the property balance sheet is stabilizing. Residential inventory and Swedish rates show whether trapped Nordic capital can recycle. The next report matters less for the absolute backlog number than for the combination of revenue conversion, margin and cash.

Cross-synthesis, conclusion, uncertainties and sources

Vertically, the capability Skanska has genuinely proved over more than a century is institutional project execution. It evolved from a Swedish concrete specialist into a multinational builder because it repeatedly accumulated references, engineers, local procurement networks and financial capacity. That history did not create immunity from bad contracts. The important recent development is that management has converted painful project experience into a more selective risk culture. A 3.5% Construction margin in 2023–2024, 4.1% in 2025 and 4.3% today is stronger evidence of that capability than the record backlog alone.

Past success came from a mixture of industrialization, post-war housing, international expansion, public infrastructure and management capability. Some of those tailwinds were era-specific. Today's data-center boom is another era tailwind. The enduring part is Skanska's ability to qualify for and deliver difficult projects across multiple cycles. That distinction matters because an investor should pay a premium for the latter and treat the former as cyclical upside.

Horizontally, Skanska sits in an attractive middle ground but does not own the richest economics in the value chain. Turner shows that a global construction manager can achieve extraordinary growth from data centers while still generating only a few percentage points of EBITDA margin. Sterling and MYR show that specialized electrical and mission-critical contractors can capture much wider margins. Skanska's opportunity is to keep an enormous book full while maintaining 4%-plus profitability. The market would be making an error if it priced the company as though AI construction converted it into a high-margin technology-infrastructure specialist.

The weakness is capital allocation across the own-development businesses. More than SEK 60 billion of capital is employed in development and Investment Properties while returns sit far below Skanska's targets. The balance sheet can afford the mismatch, but shareholder value eventually requires either higher sales/returns, lower capital employed or more aggressive recycling. Construction's negative working capital creates an unusually capital-efficient operating franchise; letting low-return development assets absorb the resulting liquidity can dilute that advantage.

I think the market is most likely underestimating the durability of the Construction improvement while simultaneously overestimating how quickly record orders become higher earnings. Those positions can coexist. The 4.3% margin is supported by multiple years of improvement and broader project selection, so I do not treat it as a one-quarter fluke. Yet 21 months of backlog means much of today's award excitement belongs to 2027–2028 revenue, and Skanska has given no evidence that the data-center contracts carry superior margins.

The 12-month variable is conversion: can quarterly Construction revenue begin to reflect the order surge without margin falling below 4%? The three-year variable is capital recycling: can RD, CPD and Investment Properties move toward target returns while the Construction book is harvested? The five-year variable is whether Skanska can institutionalize the risk discipline strongly enough that 4%-plus Construction margins survive the next downcycle rather than merely the current favorable project mix.

The extraordinary dividend should have little influence on that assessment. Ordinary payout policy already provides a reasonable shareholder return. The extra SEK 5.50 was a distribution of balance-sheet capacity, not new earnings power. Paying a high dividend can be rational when development opportunities are unattractive, but it does not make the stock cheap. In fact, the more important capital-allocation choice may be avoiding low-return new property investment rather than maximizing distributions in any single year.

The data-center customer question is the clearest information gap. Visible Q3 awards already represent more than 5% of the June backlog, while Skanska refuses, or is contractually unable, to identify several customers. Repeat-client language is positive for execution credibility; it is negative for concentration transparency. A diversified public-infrastructure book offsets part of that risk. The right valuation posture is to recognize volume visibility today and wait for margin and client-mix evidence before granting a higher Construction multiple.

The US property portfolio creates a different information gap. Management says a few completed unsold properties drove Q2's charge, but the disclosed CPD balance-sheet table does not show the remaining US completed carrying amount. Investors cannot independently calculate how much additional exposure sits behind the same assets that have already been marked. My 18% base discount and 35% conservative discount to total CPD carrying value deliberately compensate for that lack of granularity.

At the current quote, valuation neither destroys the case nor provides protection against being wrong. A roughly SEK 237 conservative SOTP, SEK 313 base and SEK 391 optimistic value bracket SEK 271.90 in an intuitively sensible way: the market is charging more than the conservative business is worth, less than the base combination of a 4.3% contractor plus partially discounted property assets, and far less than a successful margin-and-property recovery case. That is the profile of an acceptable existing holding, rather than a price at which I would take material fresh cyclical risk.

The investment becomes materially better around SEK 180–190 provided three conditions remain intact: Construction rolling margin at or above 4%, adjusted group leverage no worse than modest net debt, and no evidence that CPD carrying values require a wholesale re-mark rather than incremental impairment. At that price an investor would receive more than a 20% discount to the conservative SOTP, rather than relying on the base case to earn an adequate return.

Conversely, I would overturn the constructive part of the research if Construction falls below 4% for two consecutive quarters with project losses rather than mix explaining the decline; if free Construction working capital drops below 15% of revenue while backlog contracts; if a hyperscaler cancellation reveals that a double-digit percentage of backlog rests on one end client; or if cumulative new CPD impairments exceed roughly SEK 2 billion without offsetting realized disposal gains. Those events would indicate structural deterioration in the variables carrying the valuation.

Bull reasons:

• Construction's margin has improved from 3.5% in 2023–2024 to 4.3% R12 while backlog reached a record SEK 297.5 billion and book-to-build 114%.
• Q3 already contains SEK 15.16 billion of disclosed data-center awards, adding multi-year visibility before the quarter has been reported.
• Adjusted net cash of SEK 8.7 billion provides capacity to withstand project volatility and fund development without forced capital raising.
• Nordic residential inventory is healing: completed unsold homes fell to 326 from 516 year on year while Swedish financing conditions have eased.
• CPD and RD still report unrealized surplus values, leaving upside if transaction markets normalize without further write-downs.

Bear reasons:

• Data-center client concentration and contract form are undisclosed, so record backlog cannot be translated confidently into margin or diversified earnings.
• CPD has suffered roughly SEK 1.13 billion of write-downs across Q3 2025 and Q2 2026, indicating a repeated US valuation problem.
• More than SEK 60 billion of capital remains tied to development/property businesses earning well below management's target returns.
• The current SEK 271.90 price is above the roughly SEK 237 conservative SOTP and therefore offers no conservative-case margin of safety.
• Q2 demonstrated that a record order book does not guarantee near-term revenue conversion: revenue was flat despite 23% currency-adjusted order-booking growth.

Pre-mortem: the most plausible 50% loss path begins in 2027. One or two large North American technology clients defer new campuses as power constraints or hyperscaler capital-allocation priorities change. SEK 20–30 billion of expected new phases disappear from bookings or shift beyond 2028. At the same time, one poorly structured mega-project encounters subcontractor and schedule cost pressure. Construction margin falls from 4.3% toward 3.0–3.2%, cutting operating profit by roughly SEK 2 billion. Long US rates stay high, transaction markets remain thin and Skanska takes another SEK 2–4 billion of cumulative CPD marks. Free working capital unwinds as mobilization receipts stop growing. Investors then value normalized Construction earnings nearer 9–10 times and CPD around 50–60% of carrying value. A share price around SEK 135–160 would be plausible, roughly 40–50% below today's quote. This is a stress script, not a forecast.

A second loss script requires less drama. Construction remains profitable around 3.7–3.9%, but Nordic housing recovery stalls and US offices remain illiquid. RD and CPD capital stays trapped through 2028 while the data-center boom proves a volume event with no margin expansion. The market stops paying 17 times group earnings for the story and reverts to a mid-cycle contractor valuation while applying a 25–35% discount to development carrying values. That could move fair value toward the low SEK 200s even without a major construction disaster.

The decisive fact is that Skanska today has a better contracting franchise than its consolidated valuation metrics reveal, but the current price asks the investor to assume that improvement persists while giving limited compensation for property and concentration risk. Construction is already above its formal margin target; the next leg of value must come from proving that today's unprecedented order book does not dilute that margin and from making SEK 60-plus billion of development capital productive again.

At SEK 271.90, I would keep an existing position rather than add aggressively. The base SOTP offers roughly 15% capital upside before dividends, enough to support ownership but too little to absorb a meaningful mistake in Construction margin or CPD carrying values. The company's balance sheet and execution history reduce permanent-loss probability relative to a leveraged contractor; repeated US marks, undisclosed hyperscaler concentration and a current price above conservative value prevent a stronger call.

【Company-profile scores】

• Fundamental quality: medium
• Growth: medium
• Moat: medium
• Financial soundness: strong
• Management credibility: high
• Valuation attractiveness: medium
• Risk level: medium
• Suitable investor type: cyclical

【Investment rating】

• Rating: Hold
• One-line thesis: Record backlog and a 4.3% Construction margin support value, but property discounts and unproven data-center margin upside limit the current margin of safety.
• Current-price classification: acceptable hold
• Whether to wait for a better price: yes. I would require SEK 180–190 with Construction margin still at least 4% and no broad CPD re-mark before establishing a high-conviction new position. Waiting risks missing continued data-center-led re-rating and forfeits the roughly 3.1% ordinary dividend yield.
• Target holding horizon: 3–5 years
• Expected annualized return: conservative about -1%; base about 8%; optimistic about 16%, modeled over three years including ordinary dividends but excluding another extraordinary payout.
• Max-loss risk: roughly 40–50% in the pre-mortem case, driven by Construction margin near 3%, a major technology-order slowdown, working-capital reversal and further CPD markdowns.
• Reassessment-trigger signals: Construction R12 margin below 4% for two consecutive quarters; free working capital below 15% of Construction revenue; quarterly CPD impairment above SEK 0.5 billion or cumulative new impairments above SEK 2 billion; RD sales rate below 50% with completed unsold inventory above 450; evidence that one technology end client represents a double-digit percentage of Construction backlog.

【Ideal Buy Price】180–190 SEK

Basis: this is at least 20% below the roughly SEK 237 conservative SOTP, creating a genuine discount even if Construction normalizes to 3.8% and CPD is valued at only 65% of carrying amount.

Acceptable hold price: 270–340 SEK. This brackets the base SOTP of roughly SEK 313 and remains within roughly ±15% of that value.

Clearly overvalued price: 430–450 SEK. This begins about 10% above the roughly SEK 391 optimistic SOTP and would require economics superior even to the scenario that assumes a 4.7% Construction margin and broad development recovery.

【Valuation Range】

• current: 271.90 (close as of 2026-09-04)
• bear (conservative · ideal buy zone): [180, 190]
• base (fair · acceptable hold zone): [270, 340]
• bull (optimistic · above the clearly-overvalued line): [430, 450]

The research has four material blind spots. Skanska does not disclose data-center backlog as a category, ultimate client concentration among anonymous awards, or contract form/margin for those projects. Q2 CPD disclosure does not isolate the carrying value of remaining completed unsold US assets. A fully consistent five-year owner-cash-flow series could not be reconstructed from the primary extracts retrieved without mixing segment and IFRS cash-flow definitions. Finally, the internal library reports named in the task card were unavailable as source documents, so their peer judgments could not be audited; all peer conclusions here are independent.

The primary source spine is Skanska's Q2 2026 interim report, including the cash-flow, order, development-portfolio and balance-sheet tables. The historical financial anchor is Skanska's 2025 annual reporting and its official ten-year share/dividend series. Governance and capital-allocation conclusions use the 2026 AGM communiqué, notice and current registered-share disclosures. The Q3 technology-order analysis uses Skanska's official order-booking ledger rather than announcement-date inference. Monetary and mortgage context comes from Sveriges Riksbank and Statistics Sweden data. Peer operating comparisons use primary releases from ACS/Hochtief, Sterling Infrastructure, MasTec, MYR Group and Peab. The latest Class B closing price was cross-checked against dated market-price history for September 4, 2026.

Other tickers mentioned

ACS.MC: parent of Hochtief and the closest large-scale European reference for Turner's North American construction-management and data-center expansion.

HOT.XETRA: owner of Turner and a useful benchmark for record backlog, book-to-build and digital-infrastructure order growth.

STRL.US: US specialty contractor showing the much higher margins available in mission-critical site and electrical infrastructure.

MTZ.US: US infrastructure contractor used to contrast broad power, communications and clean-energy backlog with Skanska's general contracting.

MYRG.US: electrical and utility specialty contractor illustrating where higher-margin power and commercial-industrial scopes sit in the data-center supply chain.

PEAB-B.ST: closest discussed Nordic contractor-development comparison, with record backlog and material property exposure.

NCC-B.ST: Nordic contractor reference for regional order-cycle and execution comparisons.

FER.US: concession-led infrastructure company used as a contrast to contractor economics rather than a direct valuation peer.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.