Sandvik AB (publ): The Engineering Titan's Digital & Mining Transformation

I. Introduction & Episode Roadmap (0:00 - 10:00)

On the morning of July 17, 2026, Sandvik AB published the best quarterly numbers in its 164-year history. Revenues of SEK 36.75 billion, up 24 percent year on year. Adjusted EBITA of SEK 8.31 billion — a margin of 22.6 percent, comfortably above the top of the company's own long-term target range. Order intake of SEK 37.8 billion, the fifth consecutive quarter of double-digit organic growth. Earnings per share up 63 percent.¹

The stock fell nine percent.²

That single trading day contains almost everything a long-term investor needs to understand about Sandvik today. This is a company whose reported profits have never looked better, and whose investors have never been less sure what those profits actually mean. A significant chunk of the quarter's margin expansion came not from operating brilliance but from a raw-material price shock — tungsten, the grey, absurdly dense metal at the heart of every cutting tool and rock drill bit Sandvik sells, has repriced violently as China tightened export controls. Sandvik happens to own a tungsten mine in the Austrian Alps and one of the world's largest carbide recycling systems. When the price of your key input goes vertical and you are one of the few Western players who is partly self-supplied, your inventory becomes a profit centre. Management quantified that windfall in the Machining business at SEK 550 million in the quarter — a full 380 basis points of margin — and guided that it would shrink to roughly SEK 200 million in the third quarter.¹

So: is Sandvik a structurally better business than it was five years ago, or is it a well-run cyclical enjoying a commodity distortion? The honest answer is that both are true, and the work of this article is to separate them.

The financial reality today. Sandvik trades on Nasdaq Stockholm with a market capitalisation of roughly SEK 460 billion, against a 52-week range of about SEK 232 to SEK 413 per share — a stock that has nearly doubled off its lows.³ For full-year 2024, the group reported revenues of SEK 122.88 billion and adjusted EBITA of SEK 23.58 billion, a 19.2 percent margin. For 2025, the picture was superficially flat and underneath it quite different: revenues of SEK 120.68 billion, down 2 percent as reported but up 5 percent organically, with adjusted EBITA of SEK 23.31 billion and a margin that rose to 19.3 percent despite a currency headwind that stripped roughly SEK 2.5 billion out of profit.⁴⁵ Order intake in 2025 grew 11 percent organically to SEK 128.46 billion, building a backlog that has been converting into deliveries through 2026.⁴

Read those numbers correctly and they say something specific. In a year when the Swedish krona strengthened enough to erase billions of profit, Sandvik held its margin. That is not a growth story; it is a resilience story, and resilience is exactly what management spent the prior five years claiming it was building.

The core dilemma. Sandvik's installed base is enormous and physical: drill rigs underground, loaders, crushers, and billions of carbide inserts consumed and replaced in machine shops worldwide. The strategic bet of the last six years has been that this physical estate can be converted into a software-mediated relationship — that Sandvik can own the mine plan and the machining program, not just the steel that does the cutting. The question is whether the software is a genuine control point or an expensive accessory. Digital offerings generated SEK 5.5 billion of revenue in 2025 against a 2030 target of SEK 13 billion.⁶ That is a real business. It is also roughly 4.5 percent of group revenue, which means the thesis is still mostly a promise.

The roadmap. This piece walks through the 1862 founding at Sandviken and why a Bessemer furnace still matters to a 2026 income statement; the disastrous mid-2010s and the decentralisation surgery that followed; the 2022 unbundling that turned Sandvik from a conglomerate into an engineering group; the mining and rock-processing engine that produces most of the profit; the tooling and software businesses now split into two reporting units; a hard look at whether the M&A spree was value-creating or expensive; the moat, tested against Epiroc and IMC rather than asserted; management's credibility measured against its own prior promises; and finally the bull and bear cases with the specific evidence that would confirm or destroy each.

Start where the metallurgy started.

↑ Back to Top

II. Foundational History: The Bessemer Revolution in Sandviken (1862–1990s) (10:00 - 20:00)

In the late 1850s, a Swedish merchant named Göran Fredrik Göransson made a bet that most of European industry had already written off. Henry Bessemer's new process for blowing air through molten pig iron to burn out impurities — the invention that would eventually make cheap steel possible — had been publicly humiliated. Licensees across Britain had tried it and produced brittle, unusable metal. Bessemer's reputation was in tatters.

Göransson had bought rights to the process in 1857 and thought the problem was not the idea but the inputs.⁷ Swedish iron ore from the Dannemora region was unusually low in phosphorus, the very impurity that was wrecking everyone else's results. On July 18, 1858, at Edsken, he became the first person in the world to make commercially acceptable steel using the Bessemer method on an industrial scale.⁷ It was, in the most literal sense, a process-engineering victory: same technology, better understanding of the chemistry.

Four years later, on January 31, 1862, Göransson and his associates founded Högbo Stål & Jernwerks AB at Sandviken, north of Stockholm — the company that became Sandvik.⁸ The shares were introduced on the Stockholm Stock Exchange in 1901.⁸

Why the origin story is not decoration. It is tempting to skip the 19th century entirely. Resist that. Sandvik's entire competitive identity was set in its first decades and has never really changed: the company does not win by making commodity metal cheaply. It wins by understanding materials chemistry better than the customer does, and selling that understanding in the form of a consumable part that has to be replaced. By the 1860s the product range already included drill steel for rock-drilling — the direct ancestor of the rock tools business that today feeds every underground mine on earth.⁸

The carbide inflection. The single most important product decision in Sandvik's history came in 1942, when the company established the Coromant brand and moved into cemented carbide.⁸ Cemented carbide deserves a plain-English explanation, because almost everything about modern Sandvik rests on it. Take tungsten — a metal roughly twice as dense as lead with the highest melting point of any metal. Combine it with carbon to make tungsten carbide, a ceramic-hard compound. Grind that into a powder, mix it with a metallic binder (traditionally cobalt), press it into a shape, and sinter it at extreme temperature. What comes out is a material that is nearly as hard as diamond but tough enough not to shatter — hard enough to cut hardened steel, and tough enough to smash rock ten thousand times without breaking.

The commercial consequence is the important part. A cemented carbide cutting insert is a small, geometrically precise, ferociously engineered object that wears out. So does a rock drill bit. Sandvik does not sell one machine and walk away; it sells a machine and then sells the thing that machine consumes, forever. Production of cemented-carbide tools scaled through the 1950s at Gimo in Sweden, and the T-Max tool holder system pushed the model internationally in the late 1950s.⁸

The razor-and-blade structure, in industrial form. The economics of that model are underappreciated because they are boring. A mining customer buys a loader once a decade. That customer buys drill bits, picks, wear parts, filters and service hours every week for the life of the machine — and the aftermarket carries better margins and vastly better cyclical stability than the equipment itself. The same logic runs through metal cutting: machine tool builders sell the machine, Sandvik sells the inserts that the machine destroys.

The second structural choice was distribution. Coromant was built around direct sales and local technical support — engineers who show up at the machine shop, watch the part being cut, and recommend a different grade or geometry. That is expensive. It is also very hard to replicate, and it is the reason Sandvik's tooling business has historically defended price better than distributors would predict.

Securing the input. The third choice, made much later but rooted in the same instinct, was vertical integration into tungsten itself. In 2009 Sandvik acquired the Austrian producer Wolfram Bergbau und Hütten, which operates the Mittersill scheelite mine in the Alps and controls the tungsten chain from ore through powder production to recycling.⁹ Mittersill moves over 500,000 tonnes of ore a year at an average grade of about 0.25 percent tungsten trioxide.⁹ Alongside it, Sandvik runs recycling centres in Austria and India that take back used inserts and drill bits and return them to high-quality powder.⁹ Globally only about 30 to 35 percent of tungsten is recycled; Sandvik has set a target of 90 percent waste circularity by 2030.⁹

In 2009, that acquisition looked like sensible supply-chain housekeeping. In 2026, with Chinese export licensing throttling Western tungsten availability, it looks like one of the most valuable options on Sandvik's balance sheet — and we will return to exactly how valuable, and how temporary, that advantage is.

What the history means for an investor. Three durable traits emerged: deep materials-science capability that competitors cannot buy off a shelf; a consumables-heavy revenue mix that cushions capital-expenditure cycles; and a direct technical sales model that converts product knowledge into pricing power. Those traits carried Sandvik through world wars, the 1970s stagflation, and repeated commodity busts.

What they did not do was protect the company from itself.

↑ Back to Top

III. Conglomerate Bloat, Misallocated Capex, & The Rosengren Turnaround (2000–2019) (20:00 - 34:00)

By the early 2010s, Sandvik had become the thing that Swedish industrial groups periodically become: a federation of good businesses managed as though they were one bad one.

The structural problem was easy to describe and hard to fix. Sandvik was running three fundamentally different businesses under one roof. Mining equipment is a long-cycle capital goods business driven by commodity prices. Metal cutting is a short-cycle consumables business driven by industrial production. Sandvik Materials Technology — stainless steel tube, strip and special alloys — was a capital-intensive, energy-hungry, deeply cyclical steel manufacturer. These have different capital intensities, different customers, different decision speeds, and different rational owners. Held together, they produced a blended return on capital that flattered none of them.

The Faxander era. Olof Faxander became President and CEO in 2011, arriving from the steel producer SSAB.¹⁰ His answer to the sprawl was a matrix: a global organisation with centralised functions in Stockholm intended to strip out duplication across a company that had grown by accretion. The theory was defensible. The execution collided with reality.

Matrix structures work when decisions are slow and standardisation is the main value driver. Sandvik's businesses are the opposite: a mining customer with a rig down needs a decision in hours, and a machine-shop customer chooses between grades on the shop floor. Adding reporting lines to that added latency. Meanwhile, the commodity downturn from 2012 to 2015 hit mining capital expenditure hard at precisely the moment Sandvik's fixed-cost base was least flexible, and capital continued to flow toward the steel-adjacent businesses whose through-cycle returns did not justify it.

On August 10, 2015, the board removed Faxander. The language in the announcement was studiously polite — Chairman Johan Molin described the change as "a next step in Sandvik's further development" and credited Faxander with managing the company through challenging market conditions — but the substance was a change of direction, not a change of pace.¹⁰ CFO Mats Backman served as acting chief executive during the transition.¹⁰

Enter Björn Rosengren. The board's replacement was, in hindsight, an unusually precise choice. Björn Rosengren was then CEO of Wärtsilä, and before that had spent roughly thirteen years at Atlas Copco — the Swedish industrial group whose decentralised, division-led operating model is close to a national religion in Nordic capital goods.¹⁰ He took over on November 1, 2015, at age 56.¹⁰

Rosengren's method was not subtle and did not need to be. He dismantled the matrix. Corporate overhead was cut. Full profit-and-loss accountability was pushed down into individual divisions, each of which was expected to know its own market, own its own margin, and answer for its own returns. Sandvik's board explicitly endorsed the shift, stating that it viewed the drive toward decentralisation positively and that business decisions should be taken and implemented close to the customer, in the market.¹⁰

Three things followed from that, and they compound.

First, transparency. Once a division has its own P&L, you can see which ones are structurally good and which are being carried. The modern Sandvik describes itself as a decentralised group of 23 world-leading divisions, with an explicit portfolio rule that a business should be number one or number two in its chosen market or segment.⁶ That rule is only enforceable if you can measure each unit separately.

Second, speed. Decentralisation is often sold as a cost programme; its real payoff is decision latency. A division president who can approve a price increase or a capacity investment without a Stockholm committee responds to a downturn in weeks rather than quarters. Sandvik's ability to protect margin through the 2025 currency shock is, at least in part, a legacy of that architecture.

Third, and most consequentially, portfolio logic. If you can see clearly that a business will never earn its cost of capital inside the group, the honest conclusion is that it belongs somewhere else.

The pre-work for the breakup. On May 6, 2019, Sandvik's board announced it had decided to initiate an internal separation of Sandvik Materials Technology, and authorised exploration of a possible separate listing on Nasdaq Stockholm. Rosengren's framing was that separation would "allow full focus on Sandvik Materials Technology's key strengths and its further improved performance." Molin, still chairman, put the shareholder argument more directly: each part would develop more favourably by itself, increasing opportunities for profitable growth and improving long-term shareholder value.¹¹

Note the sequencing, because it matters for judging management. The decentralisation came first; the divestment logic followed from it. That is the correct order. Companies that announce breakups before they have divisional accountability usually discover, expensively, that they cannot actually separate the plumbing.

Rosengren left for ABB in 2020 without executing the separation himself. The task — and the harder task of deciding what to do with the proceeds of a simplified company — passed to his successor.

↑ Back to Top

IV. The Great Unbundling: The 2022 Alleima Spin-Off & Portfolio Cleansing (34:00 - 44:00)

There is a particular kind of asset that quietly taxes a diversified industrial group. It is not loss-making. It is not embarrassing. It simply consumes capital at a rate its returns do not justify, and — worse — it teaches the market to value the whole company as though every part were like it.

Sandvik Materials Technology was that asset. Advanced stainless steels and special alloys are genuinely sophisticated products with real technical barriers. They are also made in furnaces, and furnaces require enormous fixed investment, consume enormous energy, and swing with the oil-and-gas and process-industry cycle. Inside a group trying to argue that it was an industrial technology company with software attached, a steel mill was a narrative problem as much as a financial one.

The separation, executed. Sandvik's shareholders approved the distribution and listing on April 27, 2022.¹² The mechanics used Sweden's Lex Asea rules for tax-efficient spin-offs: shareholders received one share in the new company for every five Sandvik shares held. The business was renamed Alleima AB and began trading on Nasdaq Stockholm on August 31, 2022, with roughly 250.9 million shares distributed.¹²¹³

What the group actually gained. The immediate effect was on capital intensity and volatility. Removing a steel producer removed a large, lumpy, energy-exposed capital expenditure line and a chunk of earnings cyclicality that had nothing to do with mining or machining demand. It also let Sandvik raise its stated ambition: the adjusted EBITA margin target sits at 20 to 22 percent through the cycle, alongside a 7 percent annual growth target — organic plus acquisitions — and a newer capital efficiency target of net working capital at 25 percent of revenues.¹⁴¹⁵

The evidence that this was more than cosmetic sits in the return metrics. Return on capital employed rose to 15.2 percent in 2025 from 13.4 percent a year earlier, and to 16.5 percent excluding amortisation of acquisition surplus values.⁵ By the second quarter of 2026 those figures had reached 17.5 percent and 18.7 percent respectively.¹ A company that structurally shed its most capital-hungry unit and then grew into its remaining asset base is the mechanism behind that improvement; it is not a one-off accounting artefact.

The wider portfolio surgery was larger than the spin-off alone. Between 2020 and 2025, Sandvik divested or spun out businesses representing roughly SEK 30 billion of revenues while acquiring companies representing over SEK 22 billion of revenues.⁶ In other words, the group replaced roughly a quarter of its revenue base with different revenue — a portfolio rotation on a scale that is easy to miss when headline sales look flat.

Myth versus reality: the activist story. The widely repeated version of this history casts Cevian Capital, the Swedish-founded activist fund associated with Christer Gardell, as the force that pried Materials Technology loose. It is a satisfying narrative and it fits Cevian's well-documented playbook at other European industrials. But it does not survive contact with Sandvik's own disclosure. Cevian does not appear among Sandvik's ten largest shareholders in the company's published ownership tables, and Sandvik's separation announcement attributes the decision to the board's own strategic review rather than to shareholder pressure.¹¹¹⁶ Sandvik's disclosed register at the end of 2024 was led by Industrivärden at 14.6 percent, followed by Alecta at 4.5 percent, BlackRock at 3.5 percent, Swedbank Robur at 3.4 percent, Vanguard at 3.3 percent and Lundbergföretagen at 3.1 percent, across roughly 135,000 shareholders with about 40 percent of capital held outside Sweden.¹⁶

The reality is less dramatic and more instructive. Sandvik's simplification was driven by an anchor-owner governance model rather than an activist campaign. Industrivärden — the Handelsbanken-sphere holding company in which the Lundberg family is itself a major owner — is a patient, concentrated, board-represented shareholder of exactly the type that can sponsor a multi-year restructuring without demanding a quarterly scalp. Helena Stjernholm, Industrivärden's chief executive, sits on Sandvik's board.¹⁷

That distinction matters for forecasting behaviour. An activist-driven breakup tends to end when the activist exits. An anchor-owner-driven simplification tends to continue, slowly, as long as the logic holds — which is roughly what has happened, right through to the January 2026 decision to split the machining business in two.

The uncomfortable question. Did shareholders actually capture the value? The spin-off removed a low-return business, but it did not remove cyclicality — it concentrated the remaining cyclicality into mining and industrial production. Sandvik today is more exposed to two capital cycles, not fewer. The bet is that mining aftermarket and tooling consumables are stable enough to carry the group through the equipment downswings. The last two years have supported that bet. A genuine mining capex bust has not yet tested it.

Which brings us to the engine room.

↑ Back to Top

V. Core Engine Deep Dive: Mining, Rock Solutions & Processing (SMR & SRP) (44:00 - 62:00)

Picture a working face two kilometres below the surface of a copper mine in Chile or Zambia. It is hot — rock temperature rises roughly 25 degrees Celsius per kilometre of depth — and every cubic metre of air down there has been pushed from the surface by ventilation fans that consume a punishing share of the mine's total electricity. A diesel loader working that face produces heat and exhaust that the ventilation system must then remove, which is why underground diesel is not merely an emissions problem but an enormous, permanent operating cost.

This environment is Sandvik's most profitable habitat.

The scale of the mining business. In 2024, the business then named Sandvik Mining and Rock Solutions generated revenues of SEK 63.61 billion — about 52 percent of the group — and adjusted EBITA of SEK 12.95 billion, a 20.4 percent margin.⁴ In 2025 the segment delivered revenues of SEK 62.97 billion with adjusted EBITA of SEK 13.05 billion, lifting the margin to 20.7 percent even as reported revenue fell on currency translation.⁵ Order intake grew to SEK 69.2 billion.⁵

Then 2026 happened. In the second quarter, the renamed Mining business area took more than SEK 20 billion of orders in a single quarter for the first time, with revenues of SEK 18.53 billion and adjusted EBITA of SEK 3.79 billion at a 20.5 percent margin.¹

Where the profit actually comes from. The composition of that order book is more revealing than its size. In the second quarter of 2026, aftermarket organic order intake grew 17 percent while equipment orders grew just 2 percent against tough comparables, and aftermarket accounted for 66 percent of segment revenues against 34 percent for equipment.¹ Roughly two-thirds of Sandvik's mining revenue is therefore parts, service, consumable rock tools and digital subscriptions — the recurring layer — not machines.

That mix explains a great deal about the margin. Equipment is competitively bid, tender by tender, against a small number of highly capable rivals. Aftermarket is bought by a maintenance planner who cares about uptime, has the machine already, and is not going to risk an unplanned stoppage to save a few percent on a wear part. It is the difference between winning a contract and being the default.

It also explains a great deal about the cycle. When commodity prices fall and miners defer new machines, the existing fleet keeps grinding rock and keeps consuming parts. Sandvik's own second-quarter breakdown showed 66 percent of mining orders coming from brownfield expansions rather than greenfield projects — production at existing mines, not speculative new ones.¹ Brownfield demand is materially less volatile than greenfield.

The equipment portfolio in plain terms. Sandvik builds the machines that break, load and move rock underground: drill rigs that bore blast holes into a face; loaders — LHDs, for load-haul-dump — that scoop broken ore; underground trucks; surface drill rigs for open-pit benches; and, since 2021, the ground support that stops the roof falling in. That last piece came from DSI Underground, acquired for approximately EUR 943 million on a cash and debt-free basis and completed on July 7, 2021.¹⁸ DSI brought roughly EUR 516 million of 2020 revenues and about 2,000 employees in bolting and reinforcement — consumables installed continuously as a mine advances, sold through a distribution network in dozens of countries.¹⁸

Sandvik was candid at the time that DSI was dilutive to segment margin by roughly 200 basis points including purchase price allocation.¹⁸ That honesty is worth noting: the company bought recurring revenue and volume, and told investors it would cost reported margin to do it. Judging the deal therefore requires looking at aftermarket share and cash flow, not headline margin.

The war game: Sandvik versus Epiroc. No competitive analysis of Sandvik is honest without Epiroc, which was spun out of Atlas Copco in 2018 and is the closest thing to a mirror image that exists in industrial equipment. Both are Swedish. Both sell underground hard-rock equipment, rock tools, automation and service. Both have roughly two-thirds aftermarket revenue mixes. Both are chasing electrification and autonomy.

The second quarter of 2026 offers a rare clean comparison, since both reported on the same day. Epiroc booked orders of SEK 17.31 billion, up 13 percent organically, with revenues of SEK 16.70 billion, up 11 percent organically, and an adjusted operating margin of 20.1 percent against 19.7 percent a year earlier.¹⁹ Sandvik's mining orders grew 11 percent organically — 13 percent excluding major orders — with revenue growth stronger and margin at 20.5 percent.¹

Read carefully, that comparison says the two companies are performing almost identically, which is itself the finding. This is a duopoly in premium underground hard rock, and neither side is running away. Epiroc's equipment orders grew 30 percent organically in the quarter — well ahead of Sandvik's 2 percent equipment growth — which on its face suggests Epiroc won more machine placements in that specific window.¹⁹¹ One quarter is not a trend, and order timing in capital goods is lumpy. But equipment placement is the leading indicator of the next decade's aftermarket annuity, and an investor who ignores a gap that size is not paying attention.

Caterpillar and 小松製作所 Komatsu dominate open-pit mining — the giant haul trucks and shovels of surface operations — but are considerably less entrenched in narrow-vein and underground hard-rock drilling, where machine dimensions, ventilation constraints and drilling precision favour the specialists.

Rock Processing: the smaller, harder business. The crushing and screening segment is the group's structural laggard, and management has never pretended otherwise. In 2024 it produced revenues of SEK 10.70 billion and adjusted EBITA of SEK 1.56 billion at a 14.6 percent margin; in 2025, SEK 10.44 billion and SEK 1.55 billion at 14.8 percent.⁴⁵ In the second quarter of 2026, revenues of SEK 2.69 billion and adjusted EBITA of SEK 392 million held the margin at 14.6 percent.¹

Why the persistent gap of roughly six percentage points versus mining? Because crushers and screens serve two different worlds. In mining, they are mission-critical plant. In infrastructure — aggregates, demolition, recycling — they are sold through dealers into a fragmented, price-sensitive market against Metso, Terex and a long tail of regional builders. Aftermarket was 59 percent of segment revenue in the second quarter, below the mining level.¹ Less recurring revenue, more dealer intermediation, lower margin. The economics are consistent, and they are structural rather than a fixable execution problem.

Sandvik's response has been to move the segment upstream into higher-value niches. On June 22, 2026, it agreed to acquire the Italian filter press manufacturer Diemme Filtration, based in Lugo with roughly 200 employees and estimated 2026 revenues of about SEK 1.1 billion, to be reported as a new Filtration division within Rock Processing.²⁰ Filtration and dewatering address one of mining's genuinely growing problems — tailings management and water recovery — and Sandvik stated the business is margin-accretive to Rock Processing and expected to earn its cost of capital within three years.²⁰ The purchase price was not disclosed.²⁰

That non-disclosure is a recurring feature of Sandvik's M&A and a legitimate governance irritation, which the capital allocation section revisits.

Electrification and autonomy: the real switching-cost machinery. Return to that hot working face. A battery-electric loader produces no exhaust and far less heat, which means less ventilation, which means less capital expenditure on shafts and fans and less electricity forever. That is why battery-electric vehicles underground are not primarily an environmental story — they are a mine-design story, and mine designs last decades.

Sandvik booked its largest-ever BEV order in the second quarter of 2025: roughly SEK 750 million from South32 for the greenfield Hermosa critical minerals project in Arizona, comprising bolters, development drills, longhole drills, loaders and trucks, with deliveries beginning in the fourth quarter of 2026 and running through 2030.²¹ In November 2025 the company launched what it described as the first battery-electric cable bolter on the market, and in 2025 completed a fully electric tracked crushing and screening train.⁵

The automation layer sits on top. AutoMine is Sandvik's autonomous and tele-remote fleet system; in the second quarter of 2026 the company launched AutoMine Aura, a new generation using 3D perception-based navigation that Sandvik claims increased material moved by more than 15 percent in validated customer operations.¹ In the fourth quarter of 2025 it won two automation orders together worth over SEK 160 million, and launched DataDrive'31, an AI-based analytics and decision-support platform.⁵

Here is the mechanism that matters to an investor. Once a mine's traffic management, machine control and operator training are built around one vendor's autonomy stack, switching vendors means re-engineering the underground network, retraining crews, and accepting production risk in an environment where a week of lost hoisting is worth more than the entire equipment fleet's price difference. That is a genuine switching cost — and unlike a software licence, it is anchored in physical infrastructure.

The caveat is that Epiroc is executing the identical playbook, and its chief executive Helena Hedblom pointed to "encouraging adoption of our automation and digital solutions" on the same day Sandvik touted AutoMine Aura.¹⁹ Neither company is disintermediating the other. What both are doing is raising the cost for a third entrant.

↑ Back to Top

VI. Precision Tooling & The Digital Software Bet: Sandvik Manufacturing Solutions (SMM) (62:00 - 78:00)

In a job shop in Ohio or Baden-Württemberg or Ningbo, a machinist loads a block of titanium into a five-axis machining centre and presses cycle start. Over the next four hours, a spinning carbide tool removes eighty percent of that block's mass, generating a chip temperature high enough to glow, and produces an aerospace bracket with tolerances measured in microns. The tool that did the cutting cost perhaps forty dollars and is now scrap. Tomorrow the shop buys another one.

Multiply that by every factory on earth and you have Sandvik's second engine.

The numbers, and the split. In 2024, Sandvik Manufacturing and Machining Solutions generated revenues of SEK 48.57 billion — roughly 40 percent of the group — and adjusted EBITA of SEK 9.72 billion at a 20.0 percent margin.⁴ In 2025, revenues of SEK 47.27 billion and adjusted EBITA of SEK 9.39 billion held the margin at 19.9 percent.⁵

Then, on January 1, 2026, Sandvik split it in two. Announced at the May 20, 2025 Capital Markets Day, the group restructured into four business areas — Mining, Rock Processing, Machining, and Intelligent Manufacturing — explicitly to sharpen focus on profitable growth and give investors transparency on the two very different businesses that had been bundled together.¹⁴ In February 2026, Sandvik published proforma 2025 figures for each: Machining with revenues of SEK 44.00 billion and adjusted EBITA of SEK 8.70 billion at a 19.8 percent margin; Intelligent Manufacturing with revenues of SEK 3.12 billion and adjusted EBITA of SEK 686 million at a 22.0 percent margin.²²

That disclosure decision deserves credit. Bundling a SEK 3 billion software business inside a SEK 44 billion hardware business is an excellent way to avoid being held accountable for either. Separating them is an invitation to scrutiny, and management took it voluntarily.

The brands. Machining runs a portfolio of tooling brands that would each be a significant company alone: Sandvik Coromant at the premium end, Seco Tools, Walter, and Dormer Pramet serving the mid-market. That last positioning matters more than it sounds. In the second quarter of 2026 Sandvik launched a new Dormer Pramet steel turning grade, T9425, with a new coating platform and post-treatment method, described as an important launch to capture mid-market growth.¹ The premium brands defend against Iscar; the mid-market brands defend against everyone cheaper.

The competitive structure. Sandvik's principal global rival in metal cutting is IMC, the group built around Iscar and owned by Berkshire Hathaway — a privately held competitor that does not report quarterly and does not need to. Kennametal is the listed American comparator, historically operating at lower margins. Kyocera, Mitsubishi Materials and Guhring hold strong regional and specialist positions.

And then there is China. On the second-quarter 2026 call, an analyst pressed Sandvik on Chinese toolmakers competing on regional pricing. Stefan Widing's answer was direct: "We are selling on value, not price. Chinese players compete on price. Nothing new."²³ That is a defensible position and also a slightly complacent one. It is true that a shop cutting aerospace titanium buys on tool life and process reliability rather than unit cost. It is equally true that general engineering — Sandvik's single largest cutting-tool segment — is exactly where a good-enough tool at a materially lower price does erode share over time, quietly, one shop at a time. This is a claim to test through cycles, not to accept.

The tungsten shock, and what it did to the P&L. Now to the thing that made Sandvik's 2026 numbers look extraordinary.

China accounts for the large majority of global tungsten supply, and in early 2025 it imposed export controls on several tungsten products. The screw tightened through 2026: Beijing's Ministry of Commerce announced restrictions on January 6, 2026 covering dual-use exports of ammonium paratungstate, tungsten oxide and tungsten carbide, and Chinese APT exports to Japan had already fallen roughly 70 percent year on year following the February 2025 licensing regime.²⁴ APT prices, which had traded near $1,050 to $1,115 per metric tonne unit at the end of 2025, were quoted at $1,090 to $1,150 in the first week of 2026 and kept climbing.²⁴

For most Western toolmakers, this is an unmitigated cost shock. For Sandvik — with the Mittersill mine, European processing, and a recycling loop — it became something stranger. As Widing explained on the first-quarter 2026 call, Sandvik's powder business more than doubled its order intake as surging prices pulled demand and value through the chain, and cutting-tool orders grew 18 percent against revenue growth of 10 percent because customers were pre-ordering to lock in supply and price.²⁵ That gap between orders and revenues is the pre-buying, he noted — customers placing orders for later delivery.²⁵

The mechanics of the margin windfall come down to lag. As CFO Cecilia Felton described it, powder pricing lags APT prices by about one month while the cost side lags by about three months.²⁵ When prices rise fast, revenue reprices before cost of goods sold does, and the difference drops straight to EBITA. In the second quarter of 2026 that timing effect was worth an estimated SEK 550 million to Machining — 380 basis points of that segment's 28.7 percent margin, achieved on revenues of SEK 14.65 billion with organic operating leverage of 55 percent.¹

Strip that out and the underlying picture is still strong but far less spectacular. Asked on the second-quarter call what a sustainable Machining margin looks like, Widing was notably unwilling to give a clean number: "I think sustainable is a funny word. I think it still includes a premium powder margin."²³ Felton added that price versus inflation remained dilutive on cutting tools because raw-material cost pass-through was still lagging.²³ Management indicated roughly 25 percent as a reasonable sustainable underlying level, with powder representing 19 percent of Machining invoicing in the quarter.²

The analytical conclusion is uncomfortable but clear. Sandvik's tungsten integration is a real and rare structural advantage in a constrained market — but a meaningful slice of the reported 2026 margin expansion is a timing artefact that reverses when prices stabilise, and reverses against the company if prices fall. The nine percent share-price decline on results day was not investors doubting the business; it was investors marking down the quality of the earnings. That was a rational response, and the company's own guidance — the tungsten effect dropping from SEK 550 million to roughly SEK 200 million in the third quarter — confirmed the direction.¹

The software bet. Which brings us to why Sandvik went looking for a business whose margins do not depend on metal prices at all.

The strategic insight is worth stating carefully because it is genuinely clever. When an engineer designs a part, the design becomes a CAD file. Before that part can be cut, someone must write the toolpath — the instructions telling the machine where to move, how fast, how deep, and with which tool. That translation happens in CAM software. Whoever controls the CAM software sits upstream of the tool selection decision, at the exact moment it is made. The same logic holds underground: a mine plan produced in mine-planning software specifies the drilling and loading sequence that determines which equipment and consumables are needed.

Sandvik's argument is that owning that layer converts a transactional consumables relationship into a workflow relationship. In 2025, digital offerings across the group generated SEK 5.5 billion of revenue, with a stated 2030 target of SEK 13 billion.⁶

What the software business actually looks like. Intelligent Manufacturing in the second quarter of 2026 produced revenues of SEK 880 million and adjusted EBITA of SEK 198 million at a 22.5 percent margin.¹ Recurring revenue — maintenance plus subscription — was 63 percent of segment revenue, with licences at 29 percent.¹ Reported organic order growth was 7 percent; underlying growth was 9 percent, with the two-point difference caused by the transition from perpetual licences to subscription, which depresses reported revenue in the switchover years even when the business is healthier.¹

Widing has been consistent and unusually plain about how early this is. On the first-quarter call he said subscriptions are "still at a low level for us. We are at the start of this journey," describing the subscription share as in the low single digits.²⁵ He also flagged a real seasonality quirk: Intelligent Manufacturing margins are structurally lower in the first half and much stronger in the second.²⁵

The hardware-agnostic principle. The design choice that makes the strategy credible is also the one that limits its power. Sandvik's software runs on rival hardware. Mastercam programs machines from any builder; Deswik plans mines running Epiroc and Caterpillar fleets. That openness is what lets Sandvik's software achieve industry-wide installed base — and it is why the software is unlikely to be pulled out by a competitor's counter-move.

But it cuts both ways. A CAM package that is open to all tooling suppliers cannot straightforwardly force Sandvik inserts into the toolpath without destroying its own neutrality, which is precisely the asset that gives it reach. The influence is real but soft: default tool libraries, recommended cutting data, integrated ordering. It is a nudge, not a lock. Anyone modelling a step-change in Sandvik's tooling share because it owns Mastercam should show their working.

The genuine, provable value of the software is more modest and more durable: high-margin recurring revenue with a different cyclicality from cutting tools, and a live data relationship with the customer's process. Whether it eventually becomes a control point is the open question, and 2026 does not yet answer it.

↑ Back to Top

VII. Capital Deployment & M&A Benchmark: Did Sandvik Overpay? (78:00 - 90:00)

Stefan Widing took over on February 1, 2020 — weeks before the pandemic shut the industrial world down.²⁶ Born in 1977, with a master's in applied physics and electrical engineering and a bachelor's in business administration, he had spent 2001 to 2006 at Saab and 2006 to 2020 at Assa Abloy, rising to executive vice president and head of HID Global, the group's technology division.²⁶ He was, in other words, not a metallurgist. He was a software-and-electronics operator dropped into a company built on powder metallurgy.

What followed was the most acquisitive period in Sandvik's modern history — well over twenty transactions across digital manufacturing, mining software and aftermarket consumables. Three of them defined the strategy.

Mastercam: the expensive flag-plant. Sandvik announced the acquisition of US-based CNC Software, creator of Mastercam, on August 25, 2021, and completed it on September 30, 2021.²⁷ Sandvik did not disclose the price; Reuters reported it at approximately $1.2 billion.²⁸ What Sandvik did disclose was the target's financial shape: 2020 revenues of about USD 60 million, an EBIT margin of 25 to 30 percent, and roughly 60 percent recurring revenue.²⁷

Do the arithmetic that Sandvik declined to do publicly. At $60 million of revenue and, generously, a 30 percent EBIT margin, the business earned around $18 million. A $1.2 billion price implies something in the region of twenty times revenue and well north of sixty times EBIT. Even allowing for the fact that 2021 was the absolute peak of the software valuation cycle, and even allowing that Mastercam's installed base was the largest in CAM, this was an extraordinarily full price for an industrial group whose group-level EV/EBITDA was a fraction of that.

The strategic case was that installed base in CAM is nearly impossible to build organically — machinists learn one package and stay with it for a career — and that Sandvik was buying the industry's largest position outright rather than fighting for it over a decade. That case is coherent. It is not a refutation of the price. A skeptical investor is entitled to conclude that Sandvik paid a strategic-option premium at the worst possible moment in the cycle, and that the goodwill sitting on the balance sheet from that deal is the single most impairment-exposed asset in the group.

Deswik: the better-priced twin. Announced on December 2, 2021, Deswik brought mine planning and scheduling software with rolling twelve-month revenues of AUD 79 million as of October 2021, approximately 45 percent recurring revenue, an EBITA margin of about 30 percent, and more than 10,000 licences in use.²⁹ Sandvik simultaneously created a new Digital Mining Technologies division around it.²⁹ The purchase price was again not disclosed.²⁹

Deswik is arguably the stronger of the two software deals, for a structural reason. Mine planning software sits at the centre of a mining company's engineering department and touches reserve estimation, scheduling and production commitments. The output is used to make capital allocation decisions worth billions. Software with that much organisational weight is not casually replaced. Sandvik's own reporting supports the operational thesis: Digital Mining Technologies has been repeatedly cited as a double-digit growth contributor to mining aftermarket order intake.⁵¹

DSI Underground: the unglamorous one that probably paid. Of the three, the ground support acquisition is the one that most clearly did what it promised. It added a genuinely consumable product line — bolts, cables and reinforcement installed continuously as a mine advances — with a distribution footprint Sandvik would have spent a decade building. It was margin-dilutive on day one and Sandvik said so.¹⁸ But it structurally increased the aftermarket share of the mining business, and aftermarket share is the variable that determines how much of Sandvik's mining profit survives a downturn.

Schenck Process Mining was folded into Rock Processing in 2022 to expand high-capacity screening and feeding, part of the same pattern of buying adjacent installed base rather than adjacent technology.

The disclosure problem. Sandvik has now completed three major acquisitions and numerous bolt-ons without disclosing purchase prices, including the June 2026 Diemme transaction.²⁰²⁷²⁹ This is legal, common in Swedish practice, and analytically corrosive. Without prices, outside investors cannot compute returns on invested capital for the acquisitions individually, cannot verify whether the stated hurdle — returns in line with cost of capital within three years — was actually met, and are left checking group ROCE as a proxy.²⁰ An activist would make exactly this argument: a company that acquires SEK 22 billion of revenue over five years and declines to disclose what it paid is asking shareholders to trust rather than verify.⁶

The counter-evidence is that group returns have improved rather than deteriorated through the acquisition programme, which is what one would expect if the deals were broadly sensible and would be hard to fake for six years. That is real evidence. It is not the same as deal-level accountability.

Balance sheet discipline: the part that is verifiable. Here the record is genuinely strong. Financial net debt fell to SEK 26.5 billion at the end of 2025 from SEK 32.1 billion a year earlier, with net debt to EBITDA at 0.9 times.⁵ Free operating cash flow was SEK 21.2 billion in 2025, a cash conversion of 95 percent.⁴⁵ Capital expenditure ran at just SEK 1.0 billion in the fourth quarter of 2025, about 119 percent of depreciation — a company investing enough to maintain its base without empire-building.⁵

The debt structure is conservative: total loans of SEK 33.9 billion with an average duration of 3.3 years, 84 percent long-term, an average interest rate of roughly 2.9 percent excluding swap costs, and SEK 11.2 billion of committed credit facilities.¹ Interest net is guided to about SEK -0.6 billion for 2026 against capital expenditure of SEK 4.0 to 4.5 billion.¹ For an industrial group that spent five years buying software companies, that is a notably unstressed balance sheet, and it means refinancing risk is not currently a live concern.

Shareholder returns follow a stated policy of 50 percent of adjusted earnings per share.⁶ The 2026 annual general meeting on April 28 approved a dividend of SEK 6.00 per share, up from SEK 5.75, with payment on May 6, and authorised the board to repurchase up to 10 percent of shares through the 2027 meeting.³⁰⁵ The long-term incentive plan for 2026 covers roughly 350 senior executives with three-year vesting and performance conditions, at an estimated cost of SEK 411 to 416 million.³⁰

One cash-flow flag worth watching. Free operating cash flow in the second quarter of 2026 was SEK 3.59 billion against SEK 5.09 billion a year earlier, a cash conversion of just 46 percent, driven by a SEK 4.6 billion working capital build.¹ Management attributed this to higher invoicing and, per the transcript, to inventory value rising with tungsten prices.²²⁵ That explanation is plausible and largely mechanical — you cannot grow revenue 24 percent without funding receivables and inventory. But rolling twelve-month cash conversion at 80 percent is well below the 95 percent of 2025, and working capital that builds on inflated input prices becomes a headwind if those prices reverse.² This is the single most useful number to watch in the next two quarters.

↑ Back to Top

VIII. Strategic Moat, 7 Powers & Porter's 5 Forces Analysis (90:00 - 100:00)

Strip away the narrative and ask the only question that matters over a decade: what stops someone else from taking this business?

Hamilton Helmer's 7 Powers, tested rather than asserted.

Switching costs — genuinely high, but unevenly distributed. The strongest version of this power is not the software; it is the physical automation infrastructure discussed earlier. The weaker version is the tooling relationship. A machinist who has spent years developing cutting parameters around a specific insert geometry does face real friction in switching — programs must be revalidated, cycle times re-proven, scrap risk absorbed. But that friction is measured in weeks, not years, and it collapses entirely when a purchasing director mandates a cost-down programme. The evidence that Sandvik's tooling switching costs are moderate rather than extreme is simply that its cutting-tool business is visibly cyclical and visibly price-sensitive, which a genuinely locked-in business would not be.

Scale economies — high and defensible. Sandvik invests approximately 4 percent of revenues in research and development, with an innovation sales ratio — the share of revenue from recently launched products — of 25 percent in 2025 and a target of 27 percent by 2030.⁶¹⁵ For a business whose products are consumed and repurchased continuously, that innovation cadence is a compounding advantage: each generation of coating or geometry that extends tool life by ten percent is a reason for the customer to stay. The global direct sales and service network is the second scale asset, and it is the harder one to replicate. A regional competitor can match a grade. It cannot match an engineer who arrives at the mine site in Mongolia within a day.

Process power — high, and now visibly monetised. The proprietary knowledge in carbide chemistry, powder metallurgy, and vapour deposition coatings is decades deep and largely tacit. The tungsten chain proves the point: Sandvik does not merely buy powder, it mines scheelite, processes it, and recycles used tools back into feedstock.⁹ In a market where roughly two-thirds of global tungsten is not recycled, running a closed loop is both a cost position and a supply-security position.⁹

Counter-positioning — weak, and worth being honest about. The outline's framing is that hardware-agnostic software counter-positions Sandvik against hardware-tied rivals. The trouble with calling this counter-positioning in Helmer's strict sense is that it requires the incumbent to be unable to respond without damaging its own business. Nothing prevents Epiroc, Caterpillar or a pure-play software vendor from selling open software. Autodesk, Hexagon and Siemens all sell manufacturing software that is hardware-agnostic by construction. Sandvik's software position is a solid asset acquired at a high price; it is not a structural trap for competitors.

Branding carries some weight at the premium end — Coromant genuinely commands a price premium — while cornered resource is the interesting borderline case. Mittersill is not a cornered resource in the classic sense, because it supplies only a portion of Sandvik's needs; management indicated own sources cover roughly 10 to 15 percent of requirements and pointedly declined to disclose the mine's tonnage, telling an analyst it was an operational number the company simply does not publish.²⁵ That refusal is a small but real disclosure friction: the analyst on that call noted, fairly, that the omission distorts growth analysis in a year when tungsten is moving the numbers.²⁵ Network economies are essentially absent.

Porter's Five Forces, applied.

Threat of new entrants: very low. Building an underground drill rig requires safety certification, global service coverage, and decades of application knowledge. Building carbide tooling requires a powder metallurgy plant, coating capability, and tungsten supply — the last of which is now a geopolitical constraint rather than a purchasing decision. The 2026 tungsten squeeze has, ironically, raised entry barriers for everyone outside China while handing Chinese producers a domestic cost advantage.

Bargaining power of buyers: moderate, and asymmetric. Mining majors run sophisticated procurement and can and do force price concessions on equipment tenders. But their leverage collapses in the aftermarket, where an unplanned stoppage costs far more than a parts discount saves. The evidence sits in the numbers: mining aftermarket orders grew 17 percent in the second quarter of 2026 while equipment grew 2 percent, and margins held.¹ In cutting tools, buyer power is meaningfully higher, particularly in general engineering.

Bargaining power of suppliers: rising, but partly internalised. This is the force that has changed most since 2024. Tungsten supply is now controlled by Chinese export licensing.²⁴ Sandvik's partial self-supply and recycling mitigate this better than any Western peer, but do not eliminate it.

Threat of substitutes: low. Rock must be mechanically broken; metal must be mechanically removed. Additive manufacturing is a genuine long-term consideration — a part grown rather than cut needs no insert — but adoption remains concentrated in low-volume, high-complexity applications, and Sandvik participates in additive metallurgy itself.

Competitive rivalry: high and stable. The Sandvik–Epiroc duopoly is intensely competitive but rationally so: two disciplined players with similar cost structures and similar aftermarket incentives rarely start price wars. In tooling, rivalry against a Berkshire-owned private competitor and a growing Chinese cohort is structurally tougher.

The synthesis. Sandvik's moat is strongest exactly where the outline suggests it is — mining aftermarket and process metallurgy — and materially weaker in the two places the equity story leans hardest: software as a control point, and cutting tools as a pricing fortress. That does not make it a bad business. It makes it a business whose durable value sits underground rather than in the cloud.

↑ Back to Top

IX. Management Credibility, Governance & Risk Radar (100:00 - 108:00)

The most useful test of a management team is not what it promises. It is what it says when the quarter goes wrong.

Sandvik got such a quarter in the third of 2025. Order intake beat expectations, but revenues came in about 2 percent below consensus and adjusted EBITA missed across all three segments, with mining profit down nearly 6 percent versus estimates. The stated cause was costs tied to an enterprise resource planning transition in mining, on top of a currency headwind of roughly SEK 837 million that diluted group margin by 130 basis points.³¹ Adjusted EBITA came in at SEK 5.54 billion against an expected SEK 5.74 billion.³¹

What is notable is the shape of the explanation: a specific, identifiable, self-inflicted operational cause, named without hedging, in a quarter where the easy move would have been to blame the currency alone. Management guided to a larger fourth-quarter currency headwind at the same time — roughly SEK 1 billion against SEK 800 million previously expected — rather than letting the market discover it later.³¹ Analysts, per Morgan Stanley's read, called the print mixed; the shares rose anyway on the order strength.³¹

The same pattern showed up again in the first quarter of 2026, when Rock Processing delivered a 12 percent margin against 15.1 percent a year earlier. Widing did not soften it: profitability "was weak in the quarter," driven by delivery timing on higher-margin products producing negative volume and negative mix simultaneously. "Not happy with the margin in the quarter, but we expect them to recover throughout the year," he said.²⁵ That is the language of someone describing a problem rather than managing a narrative.

Widing's record against his own targets. The five-year "Shift" strategy period from 2021 to 2025 ended with organic order intake up 11 percent and organic revenues up 5 percent for 2025, a margin of 19.3 percent, and cash conversion of 95 percent.⁵ Against a 20 to 22 percent margin target, 19.3 percent is a miss — though a miss delivered into a 130 basis point currency drag, which is the kind of context that deserves to be weighed rather than dismissed. Against the 7 percent growth target, the five-year record of 6 percent compound revenue growth from 2019 to 2025 is slightly short.⁶

So: consistent execution, honest reporting, and modest under-delivery against stated ambitions. The narrative has also been notably stable — the same framework of decentralisation, aftermarket expansion, digital growth and portfolio discipline has run from Rosengren's era through Widing's without unexplained pivots. The one significant structural change, splitting Machining and Intelligent Manufacturing, was announced a full seven months before it took effect and accompanied by proforma history.¹⁴²² That is how a management team that expects to be measured behaves.

Widing held 112,753 Sandvik shares as of December 31, 2025 — meaningful personal exposure, though not the kind of stake that would dominate his net worth.²⁶ CFO Cecilia Felton has maintained conservative leverage throughout the acquisition programme and delivers guidance with unusual specificity, including quarter-ahead currency and tungsten timing estimates that are, notably, falsifiable within ninety days.¹

Governance. Sandvik has a single share class with one vote per share — no dual-class structure, so voting power tracks economic ownership. Johan Molin was re-elected chairman at the 2026 meeting alongside a board including Claes Boustedt, Marika Fredriksson, Andreas Nordbrandt, Susanna Schneeberger, Helena Stjernholm, Kai Wärn and Widing himself.³⁰ Molin's own framing at the meeting was characteristically Swedish in its understatement: he cited the company's expertise and innovative drive and said he held a very positive view of the group's continued development.³⁰

The anchor-owner model has clear benefits and one clear cost. Benefit: patient capital that tolerates multi-year restructuring and resists value-destroying mega-mergers. Cost: a board that is unlikely to force uncomfortable questions about acquisition pricing, because the dominant owner is philosophically aligned with management. The unfinished activist agenda at Sandvik is not portfolio simplification — that work is largely done. It is disclosure.

The risk radar, mechanism by mechanism.

Tungsten reversal is the most immediate risk, and it is a two-sided one. The timing benefit that inflated 2026 margins works in reverse if APT prices fall: revenue reprices down within roughly a month while higher-cost inventory takes about three months to clear.²⁵ Asked directly on the first-quarter call what happens if prices pull back, management acknowledged the dynamic without quantifying the downside.²⁵ Investors should model the SEK 550 million quarterly effect as reversible, not recurring.¹

Short-cycle industrial demand remains the largest volume swing factor. Cutting-tool demand in the second quarter of 2026 was broadly strong — double-digit organic order growth in general industry, aerospace and defence across every major region — but medical and electronics declined, sharply so in North America.¹ European automotive has been subdued for an extended period.⁵ A general engineering slowdown hits Sandvik's most price-competitive segment hardest.

Mining capex timing. Brownfield weighting cushions this, but a genuine commodity downturn would defer both greenfield projects and BEV fleet conversions. Sandvik's own second-quarter framing was cautious in a specific way: management described SEK 2 billion-plus quarterly large orders as rare and warned against treating them as a new baseline, and expects mining aftermarket growth to normalise toward high single digits from current double-digit rates.² That is guidance discipline rather than momentum-chasing, and it is worth crediting.

Integration and impairment risk on software goodwill. If Intelligent Manufacturing's growth stalls, the Mastercam-era goodwill becomes an accounting question rather than a strategic one. The segment's second-quarter organic order growth of 7 percent — 9 percent underlying — is respectable but not the trajectory that justifies a peak-cycle software multiple.¹

Operational execution. The mining ERP transition already cost real money in 2025, and in the second quarter of 2026 Sandvik's mining operating leverage came in at 25 percent, below its historical 30 percent, due to one-time items including litigation costs and a South African mine closure.² Neither is systemic, but both are reminders that a decentralised group with 23 divisions generates a steady stream of small operational surprises.

Geopolitical and supply-chain. Tariff surcharges contributed 1.4 percent to both orders and revenues in the fourth quarter of 2025 — small, but evidence that trade policy is now flowing directly through Sandvik's pricing.⁵

↑ Back to Top

X. Playbook, Key Investment KPIs & Bull vs. Bear Verdict (108:00 - 114:00)

Three lessons that travel beyond Sandvik.

Separation creates value by making capital visible, not by making it disappear. The Alleima transaction did not conjure value from nothing. It removed a business whose capital intensity was structurally mismatched with the rest of the group, which allowed both the remaining company and the separated one to be capitalised and managed appropriately. The precondition was divisional accountability — you cannot cleanly separate what you cannot separately measure. Companies that announce breakups before doing that work usually discover the plumbing is welded together.

Aftermarket revenue is not a nice-to-have; it is the cyclical shock absorber that determines through-cycle valuation. The mechanism is simple: installed machines keep consuming parts whether or not new machines are being bought. Sandvik's group aftermarket share rose from 31 to 40 percent of revenue between 2020 and 2025, and that shift is arguably more important to the equity story than anything in the software portfolio.⁶

Industrial software must be hardware-agnostic to achieve reach — which necessarily limits how hard it can be used as a weapon. This is the tension at the heart of Sandvik's digital thesis, and it applies to every hardware company buying its way into software. Reach and lock-in trade off against each other.

The three KPIs that matter most.

One: organic order intake growth, split between equipment and aftermarket in Mining. This is the leading indicator, and the split is what makes it informative. Equipment orders predict tomorrow's aftermarket annuity; aftermarket orders reveal today's fleet utilisation and Sandvik's share of the customer's wallet. When those two diverge sharply — as they did in the second quarter of 2026, with 17 percent aftermarket growth against 2 percent equipment growth — the divergence itself is the signal.¹

Two: adjusted EBITA margin in Machining, explicitly adjusted for the disclosed tungsten timing effect. Sandvik quantifies this effect each quarter, which means investors can compute an underlying margin rather than accepting the headline.¹ The test is whether that underlying figure holds near the level management has indicated as sustainable once tungsten prices flatten. This single calculation separates the structural story from the commodity story.

Three: aftermarket share of Mining revenues. At 66 percent in the second quarter of 2026, this is the number that determines how much profit survives a mining downturn.¹ It is also the cleanest scorecard for whether the DSI Underground and Deswik acquisitions did what they were bought to do.

The bull case, stated with its evidence.

Sandvik sits at the intersection of two demand pools that are unusually well-supported. The energy transition requires copper, nickel and lithium in quantities that cannot be produced without deeper, more automated underground mining — precisely Sandvik's habitat — and electrification and autonomy make each new mine a more valuable, stickier customer than the last. The evidence that this is happening rather than merely hoped for: mining orders above SEK 20 billion in a single quarter, five consecutive quarters of double-digit organic order growth, and a backlog management explicitly said would convert into higher deliveries.¹⁵

Second, the tungsten position is a durable structural asset independent of the current price spike. In a world where the West is actively de-risking critical mineral supply, a European mine plus a closed-loop recycling system is worth more than its accounting value suggests, and it is not replicable on any short timeframe.⁹

Third, the financial architecture supports compounding: leverage around one times EBITDA, cash conversion that has run at or above 80 percent through the cycle, a fixed 50 percent payout policy, and a buyback authorisation.¹⁵³⁰⁶

The bear case, stated with equal seriousness.

The most serious bear argument is not that Sandvik is a bad company. It is that the market is currently valuing peak-quality earnings as though they were mid-cycle earnings. A meaningful share of 2026's margin expansion is a tungsten timing effect that management itself has told investors will shrink.¹ Machining's 28.7 percent quarterly margin is not a level anyone at Sandvik claims is sustainable.²³ Rolling cash conversion has fallen to 80 percent as working capital absorbs inflated inventory values.² Strip out the windfall and Sandvik is a high-quality industrial compounding at high single digits — a fine outcome, but not one that obviously justifies a premium to sector peers, which is where it has been trading.³¹

The second argument concerns capital allocation. Sandvik paid what appears to be a very full price for Mastercam at the top of the software cycle and has not disclosed it, nor the prices of Deswik or Diemme.²⁷²⁸²⁹²⁰ The software business it built is real, high-margin and growing — but at roughly 3 percent of group revenue and single-digit organic order growth, it is not yet growing fast enough to validate a strategic premium. The falsifiable test is straightforward: if Intelligent Manufacturing does not sustain double-digit underlying growth as the subscription transition matures, the acquisitions will have been a fair strategic idea executed at a poor price.

The third argument is competitive. Epiroc grew equipment orders 30 percent organically in the same quarter Sandvik grew them 2 percent.¹⁹¹ If that pattern persists across several quarters rather than one, it means Sandvik is losing machine placements — and machine placements are the seed corn of the aftermarket annuity that carries the entire investment case.

Where the argument actually lands. Sandvik in 2026 is a company that has done the hard structural work — simplification, decentralisation, aftermarket expansion, honest segment disclosure — and is now being handed a windfall it did not engineer, which is obscuring how good the underlying business genuinely is. The most important thing an investor can do over the next four quarters is refuse to be distracted in either direction: not to extrapolate the tungsten margin, and not to dismiss the operational improvement underneath it.

↑ Back to Top

XI. Outro & Primary Source Transcript Guidance (114:00 - 116:00)

The primary materials that most repay reading are the ones where management is answering questions it did not choose.

The first-quarter 2026 call, held on April 22, 2026, is the single most valuable document in the recent set. It contains the clearest explanation of the tungsten lag mechanism — one month on price, three months on cost — from CFO Cecilia Felton, an admission that Rock Processing's margin was weak with a specific cause named, and an exchange in which an analyst pushed unsuccessfully for disclosure of the Mittersill mine's annual tonnage.²⁵ That last exchange is instructive precisely because management declined: it shows where the disclosure line sits.

The second-quarter 2026 call, on July 17, 2026, is where the earnings-quality debate happened in real time. Analysts pressed repeatedly on whether the quarter's drivers were durable, and management's answers were notably careful — describing SEK 2 billion large orders as unusual rather than baseline, expecting mining aftermarket growth to normalise, and declining to put a clean number on a sustainable Machining margin.²²³ Reading those answers alongside the nine percent share-price reaction is the best available lesson in how a market prices earnings it does not fully trust.²

The Capital Markets Day of May 20 and 21, 2025, in Gimo, Sweden, set the current framework: the four-business-area structure, the reconfirmed targets, the new working-capital discipline, and a restructuring programme in Machining running to 2030 that will deliver SEK 1 billion of annual run-rate savings at a total cost of SEK 3 billion charged to reported EBITA.¹⁴ That last item is worth tracking on its own — a six-year restructuring charge is an accounting judgment as much as an operational plan, and the gap between adjusted and reported EBITA will remain wide because of it.

Finally, the fourth-quarter 2025 report closed the "Shift" strategy period with Widing's own summary — that strong results, proven resilience and strategic progress concluded both the year and the 2021 to 2025 strategy — and opened "Advancing to 2030."⁵ Set that self-assessment against the actual outcome of 6 percent compound growth and a 19.3 percent margin against 7 percent and 20 to 22 percent targets, and the reader has everything needed to calibrate how much weight to place on the next set of promises.⁶¹⁵

Sandvik AB (publ)

Table of Contents

Sandvik AB (publ): The Engineering Titan's Digital & Mining Transformation

I. Introduction & Episode Roadmap (0:00 - 10:00)

II. Foundational History: The Bessemer Revolution in Sandviken (1862–1990s) (10:00 - 20:00)

III. Conglomerate Bloat, Misallocated Capex, & The Rosengren Turnaround (2000–2019) (20:00 - 34:00)

IV. The Great Unbundling: The 2022 Alleima Spin-Off & Portfolio Cleansing (34:00 - 44:00)

V. Core Engine Deep Dive: Mining, Rock Solutions & Processing (SMR & SRP) (44:00 - 62:00)

VI. Precision Tooling & The Digital Software Bet: Sandvik Manufacturing Solutions (SMM) (62:00 - 78:00)

VII. Capital Deployment & M&A Benchmark: Did Sandvik Overpay? (78:00 - 90:00)

VIII. Strategic Moat, 7 Powers & Porter's 5 Forces Analysis (90:00 - 100:00)

IX. Management Credibility, Governance & Risk Radar (100:00 - 108:00)

X. Playbook, Key Investment KPIs & Bull vs. Bear Verdict (108:00 - 114:00)

XI. Outro & Primary Source Transcript Guidance (114:00 - 116:00)

References