Ependion AB (EPEN.ST) - Deep Dive Research Report

Prepared 2026-06-06. Listing venue: Nasdaq Stockholm (Mid Cap). Reporting currency: SEK. Formerly Beijer Electronics Group AB; renamed Ependion AB in 2023.

Section 1: What the Company Does

Ependion makes the rugged electronic "nervous system" that lets industrial machines be operated, connected, and kept talking to each other in places where ordinary office IT equipment would fail. Think of a commuter train rattling at 200 km/h in winter, an electrical substation buzzing with electromagnetic interference, a ship's engine room, or a water-treatment plant. In all of these, two jobs need doing reliably for decades: a human needs a screen to see and control the process (the operator interface), and the machines need a data network that never drops a packet despite vibration, heat, dust, and electrical noise. Ependion sells the hardware and software for both jobs.

The group is deliberately structured as a holding company over two autonomous operating businesses, each with its own brand, factories, sales force, and P&L. The parent, Ependion, describes its own role narrowly: monitoring, financing, and acquisitions. The two businesses are:

• Beijer Electronics - operator panels, the human-machine interface (HMI). These are the touchscreens and the software behind them that let an operator run a machine.
• Westermo - robust industrial data communication. Ethernet switches, routers, wireless nodes, and now edge-computing devices, built to railway and substation survivability standards.

The founding story explains the structure. The company began in 1981 as Beijer Electronics, a Nordic distributor of Mitsubishi Electric automation gear. In 1990 it launched its own operator panels (the MAC series), turning a reseller into a product company. It listed in Stockholm in 2000 and grew by acquisition. The single most consequential acquisition was Westermo in 2007, a Swedish industrial-networking maker. Over the following 15 years Westermo quietly became the larger, higher-margin, faster-growing half of the group, bolting on Neratec (Switzerland, 2019), Virtual Access (Ireland, 2019), Eltec (Germany, 2021), and most recently Welotec (Germany, 2025). By 2023 the company recognised that "Beijer Electronics" no longer described a group whose centre of gravity had shifted to networking, so it renamed the parent Ependion and let Beijer Electronics revert to being one of two equal business entities under it.

The core value proposition is reliability in environments where failure is expensive or dangerous. A Westermo switch on a train is certified to EN 50155 (the railway rolling-stock electronics standard) and is expected to run for the 30-year life of the train. A Beijer panel on a marine vessel must survive salt, vibration, and wide temperature swings. Customers are not buying on price; they are buying the confidence that the device will not be the thing that fails, plus the certifications that let them install it legally.

CEO Jenny Sjödahl framed the demand backdrop on the Q2 2025 call as a market that is "stable but not robust" - a useful tell that this is a steady industrial-infrastructure business, not a hyper-growth one, whose fortunes track rail, energy, and water capex cycles rather than consumer technology.

A concrete example: when Swiss train-builder Stadler builds a fleet, it needs an onboard data network connecting the driver's cab, passenger information displays, CCTV, and signalling-related systems, all surviving vibration and temperature extremes for decades. Westermo supplies the EN 50155-certified Viper switches and Oryx train-to-trackside wireless nodes; Beijer can supply the operator displays. Westermo has won multi-million contracts from Stadler on exactly this basis. The sale is not a box; it is a qualified, certified, long-lived subsystem that the train-builder designs in once and rarely changes.

Section 2: Business Segments

Ependion reports two segments. They share a parent philosophy (decentralised, robust industrial electronics) but serve different buyers, sell through different channels, and earn different margins.

Westermo - Industrial Data Communication (the growth and margin engine)

What it does. Westermo designs and manufactures ruggedized networking equipment: managed and unmanaged industrial Ethernet switches, cellular and Wi-Fi/5G routers, millimetre-wave wireless nodes, Ethernet extenders, and (since 2025) edge-computing platforms. Its end markets are rail (both onboard rolling stock and trackside), energy (substation automation and smart grid), water and wastewater, maritime, and increasingly defence. Westermo is the larger entity - roughly 62% of group sales (group ~SEK 2.2bn in 2025, Beijer ~SEK 858m, leaving Westermo the balance).

The core capability. Two things took years to build and would be hard to replicate. First, the certification estate: EN 50155 for rail, DNV approval for marine, IEC 61850 for substations. Each certification is a multi-year, expensive process that a new entrant cannot shortcut, and customers will not buy uncertified gear for safety-critical use. Second, the software stack: Westermo's WeOS operating system provides the redundancy and network-management features (fast failover protocols) that keep a network alive when a cable is cut, which is the entire point in a moving train or a power substation.

Why it is separate. Westermo was an acquisition (2007) with its own brand equity in rail and energy, its own engineering culture, and a customer base (network architects, system integrators) entirely distinct from the machine-builders who buy HMIs. Keeping it autonomous preserved the brand and the talent.

Competitive position. Westermo competes against Moxa (Taiwan), Belden's Hirschmann and GarrettCom lines, Siemens (Scalance and Ruggedcom), Cisco's industrial Ethernet range, Red Lion (Sixnet), Phoenix Contact, and smaller players like Antaira and EtherWAN. Westermo wins on rail and harsh-environment specialisation and EMC tolerance; it is smaller than Moxa or the Siemens/Cisco giants in the broad industrial-switch market, so it competes by being the specialist the generalists are not.

How management talks about it. As the priority growth platform. The Welotec acquisition, the RazorSecure cybersecurity minority stake, and the defence push are all Westermo-centred. Its EBIT margin (mid-teens, e.g. 16.6% in Q1 2026) is roughly double Beijer's.

Beijer Electronics - Human-Machine Interface (the heritage business in transition)

What it does. Beijer designs operator panels (HMIs) - the touchscreens and the iX software behind them that let a person monitor and control a machine or process. End markets include marine, EV charging, machine-builders in discrete manufacturing, and other harsh-environment industrial applications. It is ~38% of group sales (SEK 858m in 2025).

The core capability. The combination of rugged hardware with the iX HMI software platform, plus a direct sales organisation in core markets backed by a distributor network elsewhere. The software is the stickier part: once a machine-builder has developed its control screens in iX, switching to a rival platform means re-engineering all of that.

Why it is separate. It is the original company, with a customer base of machine-builders and a distribution model distinct from Westermo's project-based network sales. Its economics are also different: lower margin, shorter order book (1-3 months versus Westermo's 6-12 months), and more exposure to the general industrial-automation cycle.

Competitive position. This is the more crowded, more competitive arena. Beijer competes against Siemens (SIMATIC HMI bundled with TIA Portal and Siemens PLCs), Schneider Electric's Pro-face, Mitsubishi (its own former principal), Rockwell, Advantech, Weintek, and value players like Maple Systems. The structural challenge is that Siemens and Schneider bundle HMI into a full automation suite (PLC + drives + HMI + software) and lock customers in at the ecosystem level, whereas Beijer is a focused HMI specialist that must win on product quality, openness (its panels talk to many PLC brands), and ruggedness rather than ecosystem lock-in.

How management talks about it. As a business being re-based around the new X3 platform. The prior X2 generation is being phased out (which dragged reported sales in 2025), and the X3 launch is the bet to re-accelerate. Management frames 2026 as a transition year: ~120 customers were evaluating X3 by end-2025, with a "gradual shift during 2026" into volume.

Section 3: Products and Business Detail

Westermo product catalogue. Westermo organises its range into animal-named product lines, each tuned to a use case:

• Viper - EN 50155-approved Ethernet switches for trains and rolling stock. The flagship rail product.
• CyBox - EN 50155 network and wireless solutions for rail and automotive onboard use.
• Oryx - millimetre-wave wireless nodes for multi-gigabit train-to-trackside and inter-carriage links (the high-bandwidth wireless that lets a moving train offload CCTV and telemetry).
• RedFox - Layer 3 rackmount switches with full routing performance for data-centre and facility networking.
• Merlin - compact cellular routers for substation automation and rail trackside.
• Ibex - rugged WLAN access points and LTE/5G routers/bridges for extreme environments.
• Lynx / SandCat - compact, low-power switches for space-constrained deployments.
• Wolverine - Ethernet extenders that build networks over long distances using existing cabling.
• RCP - modular computing platforms for high-performance data processing in extreme conditions.
• Welotec (acquired April 2025) - secure, scalable industrial edge-computing and connectivity gear, now positioned as Westermo's "technology centre for Edge Computing" and its door-opener into large grid operators.

Beijer Electronics product catalogue. The core line is operator panels: the X2 series (the prior generation, being phased out) and the X3 series (launched through 2025, the future platform), supported by the iX HMI software environment for designing the screens and logic, plus industrial PCs and related visualization hardware. The X3 is the strategic product: a modern HMI platform meant both to migrate the existing installed base off X2 and to win new customers who value an open, modern alternative to Siemens/Schneider lock-in.

Certifications and process knowledge. The hard-to-replicate moat is regulatory: EN 50155 (rail electronics), DNV (marine), IEC 61850 (substation). These require designing for extreme temperature, vibration, EMC, and decade-long lifecycles, then proving it through expensive third-party testing. The software (WeOS network OS, iX HMI environment) embeds redundancy and protocol know-how built over years.

Manufacturing and geography. The group is global with ~950 employees and ~SEK 307m of annual product development spend (2025). It manufactures its own hardware (Westermo in Sweden and via acquired operations in Switzerland, Ireland, Germany; Beijer with operations including Taiwan and Turkey historically). Sales are direct in core markets and via distributors elsewhere, spanning the Nordics, the rest of Europe, the Americas, and Asia. Beijer flagged a ~10% US tariff impact in 2025, indicating meaningful US sales exposure handled through price increases.

Milestones that changed the business. First proprietary panel (MAC series, 1990) turned a distributor into a product company; Stockholm listing (2000) funded acquisition-led growth; Westermo acquisition (2007) added the networking business that became the centre of gravity; the Westermo bolt-ons (Neratec 2019, Virtual Access 2019, Eltec 2021); the 2023 rename to Ependion; the Welotec acquisition (2025, EUR 35.2m initial + up to EUR 14.8m earn-out) adding edge computing; and the RazorSecure minority stake (GBP 2.7m, 2025) adding rail cybersecurity with an option for full ownership after 2027.

Section 4: Customers

Who buys. Westermo's customers are train-builders (e.g. Stadler), rail operators and infrastructure managers, electricity grid and substation operators (e.g. the E-REDES smart-grid project), water utilities, marine/offshore operators, and increasingly defence customers. Beijer's customers are machine-builders (OEMs who embed a panel into the equipment they sell), marine equipment makers, and EV-charging hardware makers (e.g. EVgo charging solutions reference).

Who makes the decision and on what criteria. For Westermo, the buyer is typically a network architect or systems integrator inside a large infrastructure project. The decision criteria are certification (does it meet EN 50155 / DNV / IEC 61850?), proven reliability, redundancy features, and long-term supply commitment. Sales cycles are long - Westermo's order book runs 6-12 months, reflecting project-based, designed-in business. For Beijer, the buyer is a design engineer at a machine-builder choosing an HMI to standardise on; the criteria are software ease, ruggedness, openness to multiple PLC brands, and price. Beijer's order horizon is short (1-3 months), reflecting more transactional, distribution-fed demand.

Why they choose Ependion. Specific reasons, not generic: the certifications competitors lack in rail and harsh environments; the WeOS redundancy software that keeps networks alive; the openness of Beijer's panels (they talk to many PLCs, unlike a Siemens panel optimised for Siemens PLCs); and a reputation for survivability over decade-plus lifecycles.

Switching costs. High on the Westermo side, lower on the Beijer side. Once a Westermo switch is designed into a train platform or a substation standard, it is qualified and certified into that design and rarely swapped - changing it means re-engineering and re-certifying. On the Beijer side, the lock-in is the iX software investment a machine-builder has made; the hardware itself is more substitutable, which is precisely why the X2-to-X3 transition is delicate (a forced platform change is a moment when a customer could evaluate rivals).

Concentration and contract structure. Management has not flagged dangerous single-customer concentration; the business is spread across many infrastructure projects and machine-builders. Revenue predictability differs by segment: Westermo's long order book (6-12 months) and rising backlog (group backlog ~SEK 1.1bn at end-2025) give good visibility; Beijer's short book makes it more cycle-sensitive. The mix is a blend of project/contract business (Westermo, milestone-driven, multi-year) and shorter-cycle product orders (Beijer, distribution and OEM). Welotec adds a 3-6 month order horizon, sitting between the two.

Section 5: Competitive Landscape

The two segments sit in different competitive structures, so they must be assessed separately.

Westermo (industrial networking). This is a fragmented global market with a few giants and many specialists. The giants - Cisco, Siemens (Scalance/Ruggedcom), and Belden (Hirschmann/GarrettCom) - have scale and broad portfolios. The largest pure-play specialist is Moxa (Taiwan), which dominates much of the general industrial-switch volume. Then come focused players: Red Lion (Sixnet), Phoenix Contact, Antaira, EtherWAN, and Westermo itself. Westermo's strategy is not to out-scale Moxa or Siemens but to own the hardest niches - rail and extreme-EMC environments - where certifications and survivability matter more than price. It wins there on its EN 50155 / DNV estate, WeOS redundancy, and decades of rail references; it loses (or simply does not compete) in the high-volume commodity-switch segment where Moxa's scale sets the price. The Welotec and RazorSecure moves are about extending the defensible niche into edge computing and rail cybersecurity rather than fighting in the commodity centre.

Beijer Electronics (HMI). This is the tougher arena. The market is led by automation conglomerates that bundle HMI into full control suites: Siemens (SIMATIC HMI + TIA Portal, dominant in Europe on the back of its PLC share) and Schneider Electric (Pro-face), with Mitsubishi, Rockwell, and Advantech also large, and value challengers like Weintek and Maple Systems undercutting on price. Beijer is a focused, mid-sized specialist. Its structural disadvantage is that it cannot bundle - a customer already standardised on Siemens PLCs faces gravitational pull toward Siemens panels. Its counter is openness (panels that work across PLC brands), ruggedness, and the iX software experience. This is a genuinely competitive, margin-pressured segment, which is exactly why Beijer's margins (~8-11%) sit well below Westermo's mid-teens. There is no broad moat here; there is a defensible position in specific rugged/marine/open-system pockets.

Barriers to entry. High in rail networking (certification cost and time, plus the trust needed to be designed into 30-year assets), moderate in general industrial networking, and lower in HMI (where Chinese and value players have entered freely). A new entrant cannot easily replicate Westermo's certification estate or rail reference base; it can more easily attack Beijer's HMI hardware on price.

Section 6: Industry

Demand drivers. Ependion sells into long-cycle infrastructure and industrial-automation spending. The structural tailwinds are: rail electrification and digitalisation (new fleets, signalling upgrades, passenger connectivity); the energy transition (smart grids, substation automation, grid expansion to connect renewables - the explicit logic behind buying Welotec and targeting grid operators); water and wastewater infrastructure renewal; Industry 4.0 / factory digitalisation (driving HMI and connectivity demand); and, more recently, European defence rearmament, which Ependion has named as a deliberate group-wide focus.

Industry size. The HMI/operator-panel market that Beijer serves is globally worth over USD 5bn and growing ~8% a year (company framing), which Beijer also describes as heading toward ~SEK 75bn by 2027 with software outgrowing hardware. The industrial Ethernet / networking market Westermo serves is a multi-billion-dollar market populated by Cisco, Siemens, Rockwell, Beckhoff, Moxa, Belden, and Huawei among others, growing on the back of industrial digitalisation.

Position in the supply chain. Ependion is a component/subsystem supplier sitting between silicon vendors (it buys memory and processors) and the system integrators / OEMs / infrastructure owners who build the trains, grids, ships, and machines. It is exposed upstream to semiconductor supply and pricing - notably memory-chip prices, which management flagged in 2025-2026 as rising due to AI-driven demand, passed through via price increases.

Regulation. The market is shaped by certification regimes more than by government policy: EN 50155 (rail), DNV (marine), IEC 61850 (substations), plus emerging rail-cybersecurity requirements (the rationale for the RazorSecure stake). These standards are simultaneously a barrier to entry and a driver of recurring product refresh.

Cyclicality. Moderately cyclical. The Westermo/infrastructure side is more resilient and visible (long order books, public-sector and utility funding), while the Beijer/machine-builder side tracks the general industrial-automation capex cycle and is more exposed to manufacturers pausing investment in uncertain times - which is exactly what management described through 2025 ("stable but not robust," "cautious customer behaviour").

Tailwinds and headwinds. Tailwinds: rail capex, grid investment, defence, digitalisation. Headwinds: a strong Swedish krona (a recurring drag on reported results, SEK 60m of negative currency impact in Q4 2025 alone), semiconductor cost inflation, US tariffs (~10% on Beijer), and geopolitical uncertainty dampening industrial investment decisions.

Section 7: Growth Triggers

All items below are drawn from the four most recent earnings calls. Calls used: Q2 2025 (Jul 16, 2025), Q3 2025 (Oct 21, 2025), Q4 2025 (Feb 2, 2026), Q1 2026 (Apr 28, 2026).

• X3 HMI platform ramp from evaluation into volume during 2026 (Q4 2025 call, Feb 2 2026; repeated Q1 2026, Apr 28 2026). ~120 customers had ordered X3 to evaluate by end-2025; management expects existing X2 customers to migrate and new customers to be won on the modern platform.

  "A gradual shift during 2026 for X3 evaluations turning into bigger volumes" (Q4 2025 call, Feb 2 2026).

• Welotec cross-selling into large grid operators and the energy segment (Q2 2025, Jul 16 2025; repeated Q3 2025, Q4 2025, Q1 2026). Welotec is positioned as Westermo's edge-computing technology centre and a "door opener" to larger grid operators, with delayed Q4 deliveries fully cleared in Q1 2026.

  Welotec's offering is "a door opener to larger grid operators, enhancing Westermo's market position" (Q1 2026 call, Apr 28 2026).

• Defence-sector expansion across the group (Q3 2025, Oct 21 2025; repeated Q4 2025, Q1 2026). Management launched a group-wide initiative to grow in defence, noting current products already suit defence applications with existing references, and is now working to understand European-market requirements. Westermo cited strong defence-sector order intake in Q1 2026.

• Rail cybersecurity via RazorSecure (Q3 2025, Oct 21 2025). A GBP 2.7m minority stake in RazorSecure adds rail cybersecurity capability, with an option to acquire full ownership after 2027.

• Train-segment order recovery at Westermo (Q4 2025, Feb 2 2026; reinforced Q1 2026). Management called out a "strong recovery in order intake from the train segment," with Q1 2026 train-network, energy, and defence bookings all strong.

• Record order intake building backlog and visibility (Q1 2026, Apr 28 2026). Group order intake hit a record SEK 778m (+41%), book-to-bill 1.3, with Westermo order intake +57% to record levels - supporting forward sales conversion given Westermo's 6-12 month order horizon.

Section 8: Key Risks

X3 transition execution risk (high probability, moderate impact). The Beijer business is mid-platform-change. Phasing out X2 already cost reported sales in 2025, and the entire re-acceleration thesis rests on ~120 evaluators converting into volume and existing customers migrating rather than re-tendering to Siemens, Schneider, or value rivals. A forced platform change is the one moment a sticky customer is most likely to shop around. If the X3 ramp slips, Beijer's drag on the group continues.

Semiconductor cost inflation, especially memory (moderate probability, moderate impact). Management has repeatedly flagged that AI-driven demand is pushing memory-chip prices and lead times up. They are passing costs through with price increases, which works only if demand holds; if customers resist or competitors absorb costs, margins compress.

Management acknowledged "price increases in memory chips due to AI demand," stating they are raising prices to compensate "with no immediate sourcing issues" (Q4 2025 call, Feb 2 2026).

Currency (high probability, moderate drag - company-specific). This is not generic forex hand-waving. Ependion manufactures heavily in Sweden and sells globally, so a strong krona directly compresses reported sales and EBIT. The hit was explicit: SEK 60m of negative currency impact in Q4 2025, and roughly SEK 9m of EBIT erosion in Q2 2025. This is a structural, recurring headwind tied to the SEK cost base.

Tariffs and trade friction (moderate probability, moderate impact). Beijer flagged a ~10% US tariff hit in 2025. Escalating trade barriers raise costs in a price-sensitive segment and could erode US competitiveness.

Cyclical industrial demand and geopolitical hesitation (moderate probability, moderate impact). Management described 2025 demand as "stable but not robust" with cautious customers delaying investment. The Beijer machine-builder business is the most exposed; a deeper industrial downturn would hit its short-order-book revenue quickly.

Acquisition integration and earn-out risk (lower probability, moderate impact). Welotec carries an earn-out of up to EUR 14.8m tied to 2025-2026 results, and Westermo has now stepped into edge computing - a "new yet closely related" technology area. Integration missteps, earn-out disputes, or failure to realise the promised grid cross-selling would undercut the energy-growth thesis. The directed share issue used to fund Welotec also diluted existing shareholders.

Competing against bundlers in HMI (structural, ongoing). Beijer permanently competes against Siemens and Schneider, who can bundle HMI into full automation suites and lock customers in at the ecosystem level. This is a durable margin ceiling on the Beijer segment, not a one-off event.

Section 9: Walk the Talk

Four concalls used: Q2 2025 (Jul 16, 2025), Q3 2025 (Oct 21, 2025), Q4 2025 (Feb 2, 2026), Q1 2026 (Apr 28, 2026). The most recent is within 90 days of today.

Starting at Q2 2025, management's posture was candid about a soft top line. Sales fell 5% as old products were phased out and prior-period orders were thin, and CEO Jenny Sjödahl described the market as "stable but not robust." Crucially, they did not pretend the weakness away; they pointed to order intake up 16% (Westermo +22%, Beijer HMI orders +23%) as the leading indicator, and framed the just-completed Welotec acquisition as the energy-growth lever. The implicit promise: orders are turning even while sales lag, and Westermo plus Welotec will carry the group.

By Q3 2025 that promise was tracking. Sales returned to growth (+10%), order intake rose 11%, and EBIT margin improved to 11.6% from 10.3%. Management added the RazorSecure rail-cybersecurity stake and explicitly launched the group-wide defence initiative and the X3 HMI series - so the strategic agenda they had been narrating was being acted on, not just described. Profitability improving while they invested is a point in their favour.

At Q4 2025 the story got more nuanced. Order intake was up 12% (4% organic) and the company emphasised it had "strengthened its gross margins throughout the year," delivering EBIT of SEK 58.3m and free cash flow of SEK 68m. But two things stand out for credibility. First, they were honest about a SEK 60m currency headwind and about Welotec deliveries slipping out of Q4 - they flagged the slip rather than burying it. Second, they made a concrete, falsifiable forward commitment: X3 would shift "gradually during 2026" from ~120 evaluators into volume. And they backed their confidence with capital - proposing to raise the dividend sixfold, from SEK 0.25 to SEK 1.50. That is a management team putting money behind its own optimism.

"A gradual shift during 2026 for X3 evaluations turning into bigger volumes" (Q4 2025 call, Feb 2 2026).

Q1 2026 then delivered the cleanest validation. The Welotec deliveries that slipped from Q4 were "fully delivered in Q1, with no expected impact on Q2" - the slip was timing, exactly as they had said, not lost business. Order intake hit a record SEK 778m (+41%), Westermo order intake +57% to record levels, and EBITA margin jumped to 12.5% from 9.7%. The defence, energy/grid, and train themes they had been repeating since Q2-Q3 2025 all showed up as concrete order strength. The one honest blemish: free cash flow turned negative on safety-stock building and higher receivables - which they disclosed plainly rather than glossing.

The pattern across four quarters is a management team that leans conservative-to-accurate rather than promotional. They consistently led with order intake as the truth-teller while reported sales lagged, they flagged problems (currency, the Welotec delivery slip, negative FCF) before being asked, and the one specific, datable delivery promise (Welotec deliveries clearing in Q1) was kept. The still-open promise is the X3 volume ramp, which is on track in narrative but not yet proven in segment numbers (Beijer sales were still down 6% nominally in Q1 2026). Net assessment: this is management that does roughly what it says, communicates bad news early, and backs its outlook with capital (the dividend hike) - credible, with the X3 conversion the one commitment left to verify over the rest of 2026.

Section 10: Shareholder Friendliness Index

Dividends. The dividend has been volatile, reflecting the board flexing payout to fund growth. For FY2022 the company (then Beijer Electronics Group) paid SEK 0.50 per share (AGM May 2023). For FY2024 the dividend was cut to SEK 0.25 per share - a deliberately cautious level set amid uncertain markets and the need to part-fund the Welotec acquisition. For FY2025 the board proposed a sharp increase to SEK 1.50 per share (a sixfold raise, ex-date 2026-05-13), signalling confidence after a record-order-intake year and strong free cash flow (FY2025 FCF SEK 68m in Q4 alone). The FY2023 dividend could not be cleanly verified across sources and is therefore left unstated rather than estimated. The trajectory is best read as: cautious during the 2024 uncertainty and Welotec financing, then a strong reset upward for 2025.

Buybacks and dilution. Ependion has not run a meaningful share-repurchase programme over this period; the share-count story is the opposite - mild dilution to fund growth and incentives. In May 2025 the company executed a directed share issue of approximately SEK 300m (around 2.9m new shares at ~SEK 103, in which board member Johan Wester subscribed for 857,100 shares) to part-finance the Welotec acquisition, lifting the share count from roughly 29m toward the low-30m range. The group also issues C-shares and transfers own shares to service its long-term incentive programmes. Net of all this, shares outstanding have grown modestly over three years rather than shrunk.

Verdict: Neutral - the company pays a (now-rising) dividend but has been issuing shares to fund acquisitions and incentives rather than retiring them, so capital return is real but partly offset by dilution.

Section 11: Insider Activities

Source: Finansinspektionen (Swedish FSA) PDMR insider register (marknadssok.fi.se). Window: trailing 12 months to 2026-06-06. Note: the register lists "Beijer Electronics Group AB Share" and "Ependion AB Share" interchangeably (same ISIN, post-rename); both refer to EPEN.ST.

Recent transactions (most recent first):

Buys - read the signal. The notable feature of the last 12 months is a cluster of small open-market purchases by multiple insiders in late 2025 and early 2026, when the shares traded roughly SEK 96-122: CEO Jenny Sjödahl (March 2026), board members Lars Eklöf (January 2026) and Per Allmér (July 2025), and senior executive Kristine Lindberg twice (February and March 2026). The individual sums are modest relative to executive pay, so no single ticket is a thunderclap, but the breadth is the signal - the CEO, two independent directors, and a senior executive all buying on the open market within a nine-month window is cluster buying, a mildly bullish conviction signal. It is also well-timed in hindsight: those buys at ~SEK 96-106 preceded a move to ~SEK 142 by late May 2026. Board member Johan Wester's much larger SEK 88m subscription in May 2025 was participation in the Welotec-financing directed issue rather than a pure open-market vote, but it still reflects a major shareholder backing the acquisition with cash.

Sells - work out the why. The large May 2026 "disposals" by Sjödahl and Wester (90,000-100,000 shares each at ~SEK 141.60) look superficially like heavy selling, but they are the mechanical unwind of a 2023 call-option incentive program maturing: each exercised options struck at SEK 68.59 and sold the resulting shares at ~SEK 141.60, monetising a program after the stock had roughly doubled. This is incentive realisation, not a bearish view of the business. The smaller CFO disposals (3,100 and 2,804 shares) are routine tax-cover sales paired with LTIP vesting. Sjödahl's October 2025 12,318-share movement was a division of joint marital property (a spousal transfer), not a market sale.

Net assessment. Stripping out the option-program monetisation, LTIP allotments, the marital transfer, and tax-cover housekeeping, the genuine open-market signal is net buying, and it is broad-based rather than concentrated in one person - the CEO, two directors, and a senior executive all bought. The only large non-incentive transaction was a board member funding the company's growth via the directed issue. Read plainly: mildly bullish. Insiders were accumulating on the open market through the soft 2025-early-2026 patch, and the only big sells are explainable as a maturing option scheme rather than a loss of conviction.

Section 12: Scenarios

Bull case. The X3 platform converts. Through 2026 the ~120 evaluators turn into volume orders, existing X2 customers migrate rather than defect, and Beijer's long sales decline finally inflects, so the smaller segment stops being a drag and starts contributing. Meanwhile Westermo keeps compounding: the record order intake of early 2026 converts into sales given the 6-12 month order book, the Welotec edge-computing capability opens doors at large grid operators just as Europe pours money into grid expansion for the energy transition, and the deliberate defence push lands real contracts against a backdrop of European rearmament. RazorSecure proves its worth and Ependion exercises its option for full ownership after 2027, making Westermo the certified, secure, connected-and-computing supplier of choice for rail and grid infrastructure. The krona softens or stabilises, removing the currency drag, and group margins drift toward Westermo's mid-teens as the mix shifts. The decentralised model keeps acquiring well-run niche businesses. In this world Ependion is a steadily compounding industrial-infrastructure platform riding rail, grid, and defence capex with a re-energised HMI franchise alongside.

Base case. Management delivers roughly what it has guided. Westermo continues to grow and carry the group, backlog stays healthy, and the energy/defence/rail themes keep producing solid order intake. The X3 ramp happens but gradually and unevenly - Beijer's sales stabilise rather than surge, and the segment remains the lower-margin half facing constant pressure from Siemens and Schneider. Currency remains an intermittent headwind that clips reported numbers in strong-krona quarters. Welotec integrates fine and contributes to energy growth without being transformational; the earn-out is paid as results come in. Capital return stays in the "neutral" zone - a respectable, possibly rising dividend, offset by ongoing modest dilution from incentive programs. The business looks much like it does today: a reliable, mid-single-to-low-double-digit grower with one strong segment and one in steady transition, run by a credible, conservative team.

Bear case. The X3 transition stumbles. Forced off X2, a meaningful slice of Beijer's machine-builder customers re-tender and some choose Siemens, Schneider, or cheaper Asian panels, and Beijer's revenue decline deepens instead of inflecting. Simultaneously, the industrial cycle rolls over - geopolitical uncertainty and cautious capex that management already described as "stable but not robust" tip into an outright downturn, hitting Beijer's short-order-book business fast and slowing even Westermo's project pipeline. Memory-chip inflation outruns the company's ability to pass it through, compressing margins, while a persistently strong krona keeps eroding reported results. The Welotec edge-computing bet proves harder to cross-sell than hoped, the grid-operator door-opening disappoints, and the earn-out becomes a point of friction. The directed-issue dilution and incentive-driven share creation look less justified if growth stalls. In this scenario the group becomes a low-growth, margin-pressured industrial caught between giant bundlers in HMI and scale players in networking, with the defence and energy optionality slow to pay off.

Sources:

• Ependion corporate site · Business entities - Beijer Electronics · Our history · Reports & presentations
• Westermo product lines and brands · Westermo company in brief
• Name change to Ependion (Cision)
• Q1 2026 earnings: Investing.com call highlights
• Q4 2025 earnings: Yahoo Finance call highlights · FY2025 financial statement (Nasdaq)
• Q3 2025: Interim report PDF (MFN)
• Q2 2025: Yahoo Finance call highlights · Interim report (Nasdaq)
• Welotec acquisition: Ependion press release (Cision) · Westermo announcement
• Insider data: Finansinspektionen PDMR register · Ependion insider transactions page
• Dividend/AGM: 2023 AGM resolutions (Cision)
• Quartr earnings calendar: Ependion (Quartr)

A note on what I could not fully verify: the FY2023 dividend per share could not be cleanly confirmed across primary sources, so I left it unstated rather than estimate it. The PDF interim reports were image/binary-encoded and not text-extractable, so segment figures rely on the company's own quarterly summaries and the IR site rather than line-item tables. No coverage of Ependion exists on SemiAnalysis, Stratechery, or MBI Deep Dives, so Section 13 (Further Reading) is correctly omitted.