S-1 Corporation (012750.KS)

Deep Dive Research Report - 28 August 2026

A note on sources before we begin. S-1 reports on a calendar fiscal year and files quarterly. Working forward from that cadence, the most recent results release should have landed in late July 2026. It did: S-1 published its Q2 2026 results on 31 July 2026, confirmed on the company's own IR earnings archive at invest.s1.co.kr.

Critically, S-1 does not hold earnings conference calls. This is not an inference from a failed search. It is the explicit subject of one of the five demands that activist fund Flashlight Capital Partners made of the S-1 board on 23 June 2026, which included "CEO-led quarterly earnings calls" as a reform the company has yet to adopt (Newsspace, 2026; Bloter, 2026). There are therefore no transcripts to work from. Sections 7 and 9 of this report are built instead on the six most recent quarterly earnings releases and the management commentary S-1 issued alongside each of them. Those six periods are listed at the head of Section 9. Where a claim would normally be sourced to a concall, it is sourced to the dated results release instead, and I flag the difference rather than papering over it.

1. What the Company Does

If someone breaks into a convenience store in Suwon at 3am, an alarm does not ring in a police station. It rings in one of two rooms - one in Suwon, one in Daegu - where about 140 people sit in front of screens across three shifts, twenty-four hours a day. A machine has already looked at the signal and decided whether a human needs to see it. If a human does, that person can pull up the camera feed, decide whether the movement is a burglar or a stray cat, and if it is a burglar, dispatch a uniformed responder from a branch a few minutes away. The store owner pays a monthly fee for this. They have probably paid it every month for years and will keep paying it, because the alternative is thinking about their store at 3am.

That is S-1's core business, and it has been for forty-five years. The company sells the SECOM brand in Korea, and roughly one in every two Korean businesses or households using a professional security service uses S-1's (BoanNews consumer survey of 2,610 respondents, 2026). It serves approximately 850,000 customer sites.

But describing S-1 as an alarm company undersells what it has become. Roughly half of the company's revenue no longer comes from guarding things at all. It comes from running buildings: managing the boilers and lifts and cleaning and parking of office towers, and brokering, consulting on, and managing the real estate itself. The other half is security in a broad sense - the monthly-fee alarm business, plus stationed guard forces at semiconductor and display fabs, plus large project work building surveillance systems for the Korean military.

How it got here

S-1 was founded on 28 November 1977 as Korea Security Enterprise (한국경비실업), and its origin is unusual: it was set up by a group of retired police officers, and in the same year it received the first private security business licence ever issued by Korea's Ministry of Home Affairs (company history page; EVOLOG corporate history). It was renamed Korea Security Assurance in 1978.

The pivotal event came in 1980. The Samsung Group, in partnership with the Japanese security company SECOM, acquired the business. That single decision explains almost everything about the company today: the brand it sells under, the technology it started with, the customer base it inherited, and the shareholder structure that is now the subject of a public activist campaign. In 1981 the company began Korea's first system security operation - the electronic-sensor-plus-dispatch model that Japanese SECOM had pioneered. It was renamed Korea Safety Systems in 1991 and, finally, S-1 in 1995. It listed on the Korea Stock Exchange in 1996.

The milestones from there track the industrialisation of a service business. One thousand system security contracts by 1985. A nationwide online monitoring network by 1986. Ten thousand contracts by 1989. A technical research institute in 1993. One hundred thousand contracts by 1998. The Seoul integrated control centre and a disaster-recovery site in 2002. Facial recognition access control deployed at the Seoul G20 summit in 2010. A Chinese subsidiary in 2011. And in 2015, the launch of "Blue Asset", the brand under which the real estate and building management business now trades - a deliberate move out of guarding and into operating.

The most recent chapter opened on 21 November 2025, when Samsung's annual president-level reshuffle named Jung Hae-rin (정해린) as S-1's incoming CEO, replacing Namgoong Beom. Per Korean press reporting of the appointment, Jung graduated in economics from Korea University and joined Samsung Electronics in 1990, working in management administration roles including the company's support team and its Europe regional headquarters; from late 2022 he served as representative director of Samsung C&T's Resort Division and concurrently as representative director of Samsung Welstory (Newstomato; TheBell). He was formally appointed representative director in March 2026. Note what this pedigree is and is not: it is a career operator and management-administration executive from inside Samsung, not a security technologist and not an outside hire.

The value proposition

S-1 sells the removal of a recurring anxiety, priced as a monthly subscription, backed by a physical response capability that is extraordinarily expensive to build.

The technical difficulty is not in the sensor. A door contact or a motion detector is a commodity. The difficulty is in the three things wrapped around it. First, the decision layer: distinguishing a real intrusion from the roughly 2.5 million monthly signals that hit S-1's control centres, most of which are false. Second, the physical response: having a trained, uniformed, insured person who can actually be at that address within minutes, anywhere in the country, at any hour. Third, the liability: standing behind the promise, which means training, certification, and a balance sheet.

The economics of that stack are the whole story. The dispatch network is a fixed cost that scales terribly on the way up and beautifully once dense. Every additional customer in a neighbourhood where a responder already patrols is close to pure incremental margin. This is why the physical security business in Korea is a three-firm oligopoly and has been for decades, and why S-1's half-share of it has proven so durable.

What it actually looks like in operation

Take a mid-size factory. S-1 installs sensors on the perimeter and doors, cameras inside, an access control reader at the gate that reads a face, and IoT sensors on the generator, the electrical room, and the water tank. The cameras feed SVMS, S-1's smart video management system, which as of the company's July 2026 description ran fourteen proprietary detection algorithms - intrusion, fire, unsafe work practice, forklift hazard, and others (Hankyung, 1 July 2026); in a February 2026 site briefing the company described the system as monitoring seventeen threat categories, including loitering, theft and industrial safety violations (Aju Business Daily, 11 February 2026).

When the plant closes, the system arms. If a sensor trips, the signal lands in Suwon or Daegu. About 78% of incoming signals are filtered and auto-processed by AI before a human sees them; the remainder are escalated to a controller, who now has an AI agent that responds to voice commands - "show the second floor lobby camera" - and returns the relevant footage with an analysis and a suggested action, up to and including prompting the operator to call the fire service (Asia Today; Aju). If dispatch is warranted, a responder goes. Meanwhile the IoT sensors on the generator have been quietly reporting temperature and leak status the entire time, so a burst pipe on a Sunday gets a phone call rather than a Monday morning discovery.

The customer pays one monthly bill. If S-1 also cleans, parks, maintains and brokers the building, that bill gets larger and the relationship gets far harder to unwind.

"We will incorporate advanced technologies such as AI, robots, and cloud into all services... evolving into a Biz & Life Care Integrated Solution Company that cares for customers' entire lives, not just their security needs."

• Jung Hae-rin, CEO greeting, S-1 corporate website

That sentence is the strategic thesis and the strategic risk in one breath. It is a real description of where the revenue mix has already moved. It is also exactly the kind of expansive language that a company with no clear growth engine reaches for, which is precisely the criticism now being levelled at it publicly.

2. Business Segments

S-1 reports under two management divisions - Security and Infrastructure - which in the most recent quarter split revenue almost exactly in half: Security at roughly 49% and Infrastructure at roughly 51% of the total in Q2 2026 (earnings release, 31 July 2026). Underneath that, the statutory reporting structure recognises three segments - Security System Service, Building Management Service, and Call Center Service and Telemarketing (company profile per annual filing). The five operating businesses below are how the company actually runs.

2.1 Physical Security (the dispatch business)

What it does. This is the historical core and still the single largest revenue line. Customers - small shops, restaurants, offices, factories, homes, unmanned stores, schools, apartment complexes - install S-1 sensors and cameras and pay a monthly fee. When something trips, a controller assesses it and, if needed, a responder is dispatched. Geography is Korea, nationwide. Customer type spans the full range from a one-person study café to a Samsung fab's outer perimeter.

The core capability. Three things took decades. The dispatch network: service branches, sales and response staff covering the entire country, fed by what S-1 says is the only dedicated security training academy in Korea. The control centres: Suwon (built 2002) and Daegu, running dual redundancy so that a fire or earthquake at one hands monitoring instantly to the other, with about 140 controllers on three shifts. And the judgement layer: forty-five years of accumulated threat criteria and response protocols, now encoded into the AI that pre-filters 78% of the 2.5 million monthly signals. That last asset is the interesting one, because it is a data moat that only accrues to whoever has been receiving the most alarms for the longest.

Why it exists separately. It is a subscription business with a fundamentally different cost structure to everything else S-1 does: heavy upfront installation, near-zero marginal cost per additional signal, and revenue that recurs until the customer closes the shop.

Competitive position. S-1 holds roughly 50.6% of the Korean market by usage, versus SK Shieldus at 23.8% and KT Telecop at 12.3%; in unprompted recall for dispatch security specifically S-1 was named by 73.6% of respondents, and on trust it scored 60.9% against SK Shieldus's 18.8% (BoanNews 2025 scorecard). This is as close to a genuine brand moat as exists in Korean services. Where S-1 loses is at the low end of residential, where KT Telecop's bundling with telecom service and IoT pricing has been taking share - KT Telecop reached 38.3% of the household security category against S-1's 49.0%, a far narrower gap than in the overall market.

How it fits. This is the cash cow and the reason the company is stable. Management's framing is that it must be transformed rather than harvested: shifting security "from labour-intensive systems to technology-driven approaches" (Q2 2026 commentary, Bloter).

2.2 Digital and Information Security

What it does. Managed security services, cloud security, and device management, sold largely into the same corporate base as the physical business and marketed under the S-1 SESP platform. The strategic pitch is convergence: one vendor for the cameras at the door and the endpoint agents on the laptops.

The core capability. Honestly, less than in the physical business. S-1's advantage here is distribution and incumbency, not technology depth. The company's disclosed R&D spend has run materially below SK Shieldus's despite S-1 being the larger company (12.07bn won against 17.02bn won on the 2025 comparison in BoanNews), and the R&D gap shows up most obviously in cyber.

Why it exists. Cross-sell. S-1's own data showed convergence customers (buying both physical and information security) up 34.8% by mid-2022 versus pre-pandemic 2019, and that 67.5% of convergence customers were offices and factories against just 33.0% of physical-only customers (BoanNews). Translation: convergence is an enterprise upsell, not a mass-market product.

Competitive position. Weakest of S-1's businesses. SK Shieldus has the deeper cyber franchise, and the pure-play Korean information security vendors are more specialised. FCP named cybersecurity explicitly as an area where S-1 "has not produced clear results" (Newspim, 24 June 2026).

How it fits. A strategic option, not a profit engine. It exists to defend the enterprise relationship.

2.3 Building Management and Real Estate Services (Blue Asset)

What it does. The largest single line inside Infrastructure. Launched as a branded business in 2015, Blue Asset covers the full life cycle of a commercial property: acquisition consulting, property management (leasing, revenue optimisation, sale and lease consulting), facility management (mechanical plant, cleaning, parking, security), energy efficiency, and eventual disposal (Newsis, 2015; Segye, 2024).

The core capability. Two things. First, the ability to bid IFM - integrated facilities management - where a client hands over every building function to one contractor rather than assembling five. S-1 can do this because it already owns the security leg, which is the leg most facilities managers subcontract. Second, BlueScan, its smart building layer: IoT sensors on generators, electrical rooms, water tanks and pipework that detect fire, leaks, outages, gas and temperature anomalies remotely, which matters most in buildings without 24/7 engineering staff (S-1 smart building page; Hankyung, July 2026).

Why it exists separately. Different customer (a landlord or asset owner, not a shopkeeper), different sales cycle (competitive tender, multi-year), different economics (labour pass-through with thinner but very sticky margins), and a different regulatory frame.

Competitive position. Fragmented and genuinely competitive. S-1 competes against dedicated Korean property and facility managers as well as the in-house teams of large landlords. Its differentiator is the bundle, not price.

How it fits. This is the growth engine by revenue. Management repeatedly attributes Infrastructure growth to "new building management contract wins and expanded management scope", and the CEO's stated direction is to push the business further toward IFM (Q1 2026 commentary; March 2026 AGM remarks).

2.4 Integrated Security (stationed guard forces)

What it does. Where dispatch security is remote and event-driven, integrated security is a permanent on-site guard force plus the systems around it. The customers are large industrial and national facilities: semiconductor and display fabs, plants, airports, ports, marine transport systems and important national installations. S-1 also provides the consulting, engineering, procurement, installation and maintenance for these sites.

The core capability. Clearance, scale and trust. Guarding a semiconductor fab is not a guarding problem, it is an industrial-espionage problem. The customer is handing a contractor physical access to the most valuable IP in the country. That decision is made once, on the basis of institutional trust, and revisited rarely.

Why it exists. It is a headcount business, not a technology business, and its margin profile and capital intensity are unlike the subscription lines.

Competitive position. Structurally advantaged by the Samsung relationship and by S-1's history with national-facility work (the G20 summit facial recognition deployment in 2010 being the showcase). This is the least contestable business S-1 has, and also the one most exposed to wage inflation.

How it fits. Steady, related-party-heavy, and the beneficiary whenever Samsung affiliates expand capex on new fabs. S-1's building and security revenue has been observed to move with affiliate capital expenditure cycles (Dealsite).

2.5 Security SI (project systems integration)

What it does. Lumpy, tendered, project-based work: designing and building large surveillance and control systems. The flagship is the Korean military's scientific border surveillance system (과학화 경계시스템). S-1 was contracted for the GOP performance-improvement programme and completed deployment of AI-video-analysis surveillance cameras and control systems to the 22nd Division GOP and coastal units in December 2024 - the first introduction of AI video analysis into the military's border surveillance (SAT Economy). The broader improvement programme runs to 2027 with a total budget of 465.1bn won, and S-1 is one of three shortlisted bidders alongside SK Telecom and KT (BusinessPost).

The core capability. Being one of the very few Korean firms that can carry a defence-grade integration project, hold the certifications, and pass the security vetting.

Why it exists. Because the revenue recognition is completely different. This is milestone-based project revenue, not subscription, and it is the reason S-1's quarterly revenue line is noisier than the underlying business. Security SI fell 25.9% year-on-year in Q2 2026 on project timing alone (Bloter, 31 July 2026).

Competitive position. Head-to-head against the telecom majors, who have deeper network engineering benches. S-1 wins on the video analytics and physical security domain; it does not win on scale of systems integration.

How it fits. The volatility line, and the source of most of the upside surprise potential.

2.6 Call Centre and Telemarketing

Small, and worth one paragraph only for what it reveals. This segment handles inbound customer complaints and requests, and 100% of its revenue is internal related-party business (Dealsite). It is a captive shared-services function that happens to sit inside a listed company. It is not a business anyone would build standing up, and it is a small piece of evidence for the activist argument that S-1's boundaries were drawn by group convenience rather than strategic logic.

Segment summary

3. Products and Business Detail

The SECOM product catalogue

Dispatch security (무인경비 / 출동경비). The foundational product. Sensors in a protected zone detect intrusion or anomaly; the signal reaches a control centre; a responder is dispatched on a 24-hour basis. Sold as commercial SECOM and household SECOM.

SECOM Easy (세콤이지). An all-in-one IoT product aimed at single-person households and very small premises, combining a 2-megapixel IP camera with a UWB (ultra-wideband) radar sensor. The combination matters technically: video alone generates false alarms from light and shadow, and radar alone cannot tell you what it detected. Fusing the two lets the device distinguish an actual intrusion from noise before it ever escalates (ETNews).

Samsung AI Door Cam. Launched December 2025 with Samsung Electronics. A dual-camera doorbell: the upper camera captures visitor faces, the lower camera watches packages left on the floor, with lower-camera images sent to a cloud AI server that notifies the user when a parcel arrives or disappears. It integrates with Samsung's SmartThings platform, is manufactured domestically in Korea, and passes Samsung Electronics' own security standards to limit hacking and video leakage risk. S-1's 24-hour emergency dispatch is sold on top as a separate paid service (Ajunews, 22 December 2025; S-1 product page). This is the clearest example of the group's distribution advantage: an S-1 service attached to a Samsung consumer device sold through Samsung's channel.

SVMS (Smart Video Management System). The AI video analytics engine. Fourteen proprietary algorithms as described in July 2026, covering intrusion, fire, unsafe work practices and forklift hazards, with real-time alerts pushed to staff smartphones. A February 2026 briefing described seventeen monitored threat categories including violence, loitering, theft and industrial safety violations, with internally-tested accuracy above 99%.

Facial recognition access control. Deep-learning readers using paired visible-light and infrared verification to defeat spoofing, with claimed 99% accuracy through masks and in low light. Deployed at the 2010 Seoul G20 summit and the Nuclear Security Summit, and used to run a unified employee-badge system across Samsung affiliate sites in Korea and overseas.

BlueScan. IoT sensors retrofitted to building plant - generators, electrical rooms, water tanks, pipework - reporting remotely on fire, leak, outage, gas and temperature. Sold both standalone and inside the building management contract.

Assurance 24 Plus (안심24 플러스). The unmanned-premises bundle: intelligent CCTV, remote voice warning, emergency dispatch, cash-box monitoring, power outage detection, and card-reader entry that blocks minors. S-1's own crime prevention research unit found 46% of unmanned store offenders were teenagers, which is why age-gating at the door is a headline feature rather than an accessory (BoanNews).

Smart School Service. Three components: an unmanned gymnasium management system that controls doors, lighting and HVAC on schedule and, at closing time, makes voice announcements, checks for remaining occupants, locks up and triggers dispatch on intrusion; an intelligent school safety package running AI CCTV analysis for seven incident types including fire, unauthorised entry and school violence; and BlueScan for pipes and electrical plant. Deployed with the Gunpo-Uiwang Education Support Office across three Gyeonggi schools, with leak-response consulting for 60 schools in Incheon (Hankyung, 25 August 2026).

Anti-drone. Introduced late 2025. RF scanners detect drone-to-controller communications; radar tracks position, velocity and heading; EO and IR cameras feed AI analysis of flight pattern and airframe shape; frequency interference neutralises the target. The hybrid RF-plus-radar architecture is the technical differentiator against radar-only systems (Hankyung, November 2025; Etoday).

Vehicle and personal. GPS fleet tracking (company-wide rollout from 2007) for corporate, sales and logistics fleets, and personal security phones.

Certifications and process knowledge

The barriers here are administrative and cumulative rather than patent-based. S-1 holds the first private security licence issued in Korea, obtained ISO 9002 in 1997, was designated Korea's first private emergency rescue organisation in 1997, received the industry's first Consumer-Centric Management certification in 2024, and holds fire safety inspection registration dating to 1993. It also operates what it describes as Korea's only dedicated security training academy, at Cheonan since 1998. None of these individually stops a competitor. Collectively, and combined with defence-project clearance, they are why the bidder list for a military border surveillance contract has three names on it.

Delivery infrastructure

There is no factory. The production asset is the network:

• Two control centres, Suwon (2002) and Daegu, dual-redundant, ~140 controllers on three shifts, processing ~2.5 million signals monthly with ~78% AI-filtered.
• Nationwide service branches with sales and dispatch staff.
• The Cheonan training centre.
• 7,025 employees with average tenure of 11.1 years and a workforce that is 92.7% male (Saramin company data, 2026).

That tenure figure is worth sitting with. An eleven-year average in a services company is very high, and it cuts both ways: deep institutional knowledge and low churn, but also a heavily tenured, seniority-weighted cost base that is difficult to reduce and highly exposed to retirement-benefit accounting - a point that became painfully concrete in Q1 2026.

Geographies

S-1 is overwhelmingly a Korean business. International expansion began with a Chinese subsidiary in 2011, followed by a Mongolian subsidiary and branch offices in the UAE, Singapore, Australia and Iraq, built in collaboration with Samsung group construction affiliates rather than through independent market entry. The company also operates the unified employee badge system across Samsung affiliate sites domestically and abroad. This is best understood as following existing customers overseas, not as international market share. Expanding overseas market conquest was named as an expected priority of the incoming CEO at the time of his November 2025 appointment.

Milestones that changed the business

1980 (Samsung/SECOM acquisition, which created the company as it now exists), 1981 (first system security operation in Korea), 1986 (nationwide online monitoring), 1996 (listing), 2002 (Suwon integrated control centre and disaster recovery), 2015 (Blue Asset launch, the pivot into building management), December 2024 (first AI video analysis deployed into the military's border surveillance system), December 2025 (Samsung AI Door Cam).

4. Customers

Who buys

Four distinct customer populations, with almost nothing in common.

Small and medium businesses and unmanned premises. Convenience stores, restaurants, study cafés, small clinics, warehouses. Korea's unmanned store count rose from 2,250 in 2020 to more than 10,000 by 2025 (cited at the March 2026 AGM). The buyer is the owner personally. The decision takes days, not months, and is triggered by an event: a break-in nearby, an insurance requirement, a decision to run unstaffed hours. Criteria are brand trust and response speed, in that order.

Households. Apartments and single-person homes. The buyer is the resident. Increasingly the trigger is a consumer product purchase (a Samsung AI Door Cam bought through Samsung's channel) rather than a security decision, which is precisely why that partnership matters.

Large enterprises and industrial sites. Semiconductor and display fabs, plants, logistics centres, corporate campuses. The buyer is a facilities or corporate security director, typically with sign-off from procurement and often from a group-level security function. Criteria are vendor track record, clearance, incident history, and whether the vendor can cover every site the company operates. Sales cycles run months and contracts run years.

Public sector. The Ministry of National Defence, education support offices, airports, ports, nuclear facilities. Competitive tender, formal evaluation criteria, multi-year delivery. The Ministry of Land, Infrastructure and Transport has planned anti-drone deployment at major airports and ports through 2026, and the Defence Ministry at military installations.

Why they choose S-1

For the SME and household buyer, the answer is almost embarrassingly simple: trust and recall. 73.6% named S-1 first for dispatch security; 60.9% named it as most trusted, more than three times SK Shieldus's 18.8%. In a category where the product is invisible until the worst night of your life, brand is not marketing, it is the product.

For enterprises, the answer is coverage plus incumbency. If S-1 already runs the badge system at every one of your sites in Korea and abroad, adding a site to that system is a purchase order, not a procurement exercise.

For public sector, it is the demonstrated ability to deliver a defence-grade project. S-1 got the AI border surveillance work done and fielded in December 2024. That reference is the sales pitch for the follow-on programme running to 2027.

Switching costs

Higher than they look, and they compound.

The physical layer is the first anchor: sensors, cabling, cameras, readers, control panels installed on the customer's premises. Ripping that out and replacing it means downtime and capital cost for a service that costs a modest monthly fee.

The integration layer is the second, and it is much stickier. Once a customer's access control is tied to its HR system, or its badge works across forty sites in six countries, the switching decision stops being a security-vendor decision and becomes an IT project.

The bundle is the third and strongest. Once S-1 runs a building's security and its cleaning, parking, plant maintenance and leasing, leaving requires re-tendering four services at once and accepting an integration risk the facilities director does not want to own.

The inertia layer is the quietest and probably the most important. Nobody gets promoted for switching alarm providers. They get fired if the new one misses a break-in. Behavioural lock-in in this category is severe, which is why market shares in Korean physical security have barely moved in a decade.

Concentration

There are two different concentration stories here and they must be separated.

At the site level there is essentially none: approximately 850,000 customer locations means no single subscriber matters. This is the healthiest possible revenue profile.

At the group level there is real concentration. S-1 handles physical and information security and building management for major Samsung affiliates, and its revenue has been observed to move with affiliate capital expenditure. The call centre segment is 100% internal. This is not straightforwardly a risk - the relationship is forty-five years old, deeply embedded, and largely explains why S-1's enterprise business is defensible - but it is a dependency on the capital cycle of a handful of related parties, and it is exactly the dependency an activist has now called into question by proposing that Samsung sell out entirely.

Contract structure

The revenue mix is unusually favourable and unusually well-balanced:

• Subscription (dispatch security, digital security, BlueScan): monthly fees, auto-renewing, cancelled only when the customer closes or moves. High visibility, high gross margin at scale, low volatility.
• Multi-year service contracts (building management, integrated security): tendered, renewable, with contractual escalation. S-1 explicitly attributed Q1 2026 Infrastructure revenue growth in part to renewal price increases. Moderate visibility, thinner margin, labour cost pass-through the key negotiating variable.
• Project (Security SI): milestone-recognised, lumpy, tender-won. Low visibility, and the single biggest source of quarter-to-quarter revenue noise.
• Product (AI Door Cam and hardware): one-off, low-margin, but a customer acquisition channel for subscription dispatch.

The practical consequence: roughly three-quarters of the revenue base is highly predictable, and the visible quarterly volatility comes almost entirely from the smallest piece. A quarter where revenue falls because Security SI slipped 25.9% is a very different event from a quarter where subscription revenue falls, and the two should never be read the same way.

5. Competitive Landscape

The structure

Korean physical security is a stable three-firm oligopoly with a very long right tail of small guard companies that do not compete for the same contracts. By usage share among consumers surveyed in 2025: S-1 50.6%, SK Shieldus 23.8%, KT Telecop 12.3%, with the remainder scattered.

That structure has three causes. Density economics: dispatch security requires responders within minutes of every customer, so the marginal cost of serving a new customer collapses with route density and rises brutally for a subscale entrant. Trust asymmetry: customers cannot evaluate the product before failure, so they buy the name they know. Capital and licensing: guard licensing, training infrastructure, control centre redundancy and liability capacity are all fixed costs incurred before the first customer.

The named competitors

SK Shieldus (Korea, private). The number two, formerly ADT Caps, trading domestically under the CAPS brand. SK Square sold a 28.82% stake to EQT Partners for 860bn won in 2023, making the Wallenberg-linked Swedish private equity firm the controlling shareholder (MTN). SK Shieldus is the more serious cyber competitor - higher disclosed R&D spend than S-1 despite being smaller - and its "CAPS Unmanned Safety Zone" directly targets S-1's unmanned store franchise. Where S-1 wins: brand trust (60.9% versus 18.8%), dispatch density, and financial position; SK Shieldus reported a Q3 2025 net loss even while growing revenue faster than S-1. Where S-1 loses: cyber depth, and an owner with a clear mandate to grow aggressively.

KT Telecop (Korea, subsidiary of KT Corp, KRX: 030200). The fastest grower of the three, with H1 2025 revenue up 11.9% and operating profit up 84.1%. Its strategy is IoT-AI solutions for residential and small business, sold through KT's telecom relationship. It has closed most of the gap in household security specifically (38.3% versus S-1's 49.0%). Where S-1 wins: enterprise, industrial and public sector, where KT Telecop barely competes. Where S-1 loses: price-led residential and micro-business, where the telecom bundle is a structurally cheaper customer acquisition channel.

Hanwha Vision (Korea, private, Hanwha group). Not a service competitor but an important adjacent one: a major camera and video management systems manufacturer. It competes for the hardware and analytics spend inside a security project, and can partner with any of S-1's rivals.

SECOM Co., Ltd. (Japan, TSE: 9735, ~JPY 2.58 trillion market cap as of 28 August 2026, source). The most unusual entry on any competitive landscape: SECOM is simultaneously the global archetype of this business model, the source of S-1's brand and original technology, and S-1's largest shareholder at 25.65%. It is not a competitor in Korea. It is listed here because it is the benchmark for what a mature guarding-plus-monitoring franchise can become, and because its shareholding is central to the governance fight described below.

ADT Inc. (USA, NYSE: ADT, ~USD 5.45bn market cap as of 24 August 2026, source). No Korean presence, but the closest listed comparable for the residential and small-commercial monitoring model.

Securitas AB (Sweden, Nasdaq Stockholm: SECU-B, ~USD 9.77bn market cap as of May 2026, source). The global guarding leader, illustrating the labour-heavy end of the industry that S-1 is explicitly trying to move away from.

Verisure (Sweden, Nasdaq Stockholm, listed October 2025, ~EUR 16.75bn at IPO close, source). The European monitored-alarm leader and, by market capitalisation at listing, the largest alarm company in the world. Relevant as evidence of what the market will pay for a pure subscription monitoring franchise with a growth story attached.

Building management competitors. A genuinely fragmented field of Korean property and facility managers, plus the in-house teams of large landlords. No single competitor dominates, and S-1 competes here on bundling rather than on scale or price.

Barriers to entry

High for the incumbent model, and falling for a piece of it.

Genuinely high: the dispatch network (a national response capability is a multi-year, capital-heavy build), the control centre redundancy, guard licensing and training infrastructure, defence and national-facility clearance, and brand trust in a category where the buyer cannot evaluate quality in advance.

Genuinely eroding: the technology layer. AI video analytics, cloud VMS and IoT sensing are increasingly available as commodity components. A cloud camera company can now sell most of a small business's security needs without a single guard. What it cannot sell is the person who shows up at 3am - which is exactly why S-1's response network, not its algorithms, remains the actual moat, and why management's pivot to "technology-centred" operations is a double-edged strategy: it improves margins while commoditising the differentiator.

Structural shifts underway

Three worth tracking. Ownership change at the number two: EQT-controlled SK Shieldus is a private equity asset with a finite hold period and a growth mandate. Bundling from the telecom side: KT Telecop is competing on customer acquisition cost rather than capability, which works best exactly where S-1 is weakest. And the ownership fight at S-1 itself, described in Section 8 and Section 11, which is the first serious external challenge to how this company is governed in its listed history.

Where S-1 is strong and where it is exposed

Strong: brand trust, dispatch density, enterprise and public sector incumbency, the Samsung distribution channel, and a subscription base that has proven remarkably insensitive to competition.

Exposed: residential and micro-business price competition, cyber capability, an ageing and tenured cost base, near-total reliance on a single country, and - most acutely - a capital structure and shareholder register that a third party has now publicly argued should not exist in its current form.

6. Industry

What drives demand

Labour scarcity and unmanned operation. The clearest single driver. Korean unmanned retail locations grew from 2,250 in 2020 to over 10,000 by 2025, and every one of them is a premises that must be secured precisely because nobody is standing in it. Extend the logic: unmanned gyms, study cafés, laundromats, warehouses, and school facilities used out of hours by the community. Security spend is a direct function of how many hours a valuable space sits unstaffed.

Ageing building stock. 44.4% of Korean buildings are more than 30 years old, and 23.7% of Korean schools exceed 40 years. Old buildings leak, their electrical plant fails, and they need monitoring and professional management. This drives both the BlueScan IoT layer and the facilities management business.

AI as a capability step-change. Video surveillance existed for decades as a forensic tool: you looked at the tape after the fact. AI analytics turns it into a preventive one. This changes who buys and what they will pay, and it is the reason the integrated security services category is the fastest-growing part of the Korean market.

Public safety and national security policy. Border surveillance modernisation, drone threat response at airports, ports and nuclear facilities, and school safety mandates. This is budget-driven demand with multi-year visibility and formal tender processes.

Size and growth

Korea's physical security market is forecast to reach 6.60 trillion won in 2026, up 12.2% from 5.88 trillion won in 2025 (itself up 9.7%), per the 2026 Security Market Whitepaper (BoanNews). The composition matters more than the headline:

• Integrated security services and smart cities: 3.08 trillion won in 2026, growing 15.9% - the largest and fastest-growing block, driven by centralised control centres and public infrastructure.
• Access control and biometrics: 1.57 trillion won, growing 15.6% - contactless authentication going mainstream.
• CCTV and video surveillance: 1.95 trillion won, growing only 4.4% - the hardware is maturing while the analytics on top of it are not.

The earlier edition of the same whitepaper put the total Korean security market (physical plus cyber) at 7.46 trillion won in 2023, of which physical security was 4.81 trillion won, growing 10.3% year on year (BoanNews).

Independent forecasters put the Korean physical security market on roughly a 6.7% CAGR to 2033 (Spherical Insights) - materially below the domestic whitepaper's near-term growth figures, which is worth noting: the two are measuring different scopes and the truth is probably between them.

The adjacent anti-drone market is forecast to grow from approximately 1.94 trillion won globally in 2021 to 17.47 trillion won by 2030 per Precedence Research, a more than nine-fold expansion.

Where S-1 sits in the chain

S-1 is a service integrator and operator, not a manufacturer. It buys cameras and sensors (from Hanwha Vision and others), builds the analytics and management software layer (SVMS, BlueScan), operates the control centres, employs the responders, and holds the customer relationship and the billing. This is a good place to sit: hardware is deflating in price and commoditising, while the recurring service layer captures the customer relationship and the margin. It also means S-1 does not capture the upside when a camera generation gets dramatically better - it just passes the improvement through.

Import dynamics

Not a meaningful frame for this industry. The service is inherently domestic; you cannot import a responder. The hardware layer is globally sourced with strong Korean domestic manufacturing (Hanwha Vision), and the Samsung AI Door Cam is explicitly domestically manufactured, with the security-of-supply argument (passing Samsung's own standards to limit hacking and video leakage) used as a selling point against cheaper imported cameras. If anything, the trend runs the other way: heightened sensitivity to foreign-made surveillance hardware is a mild tailwind for domestic vendors.

Regulation

Private security in Korea is licensed. S-1 has held licence number one since 1977. Beyond licensing: fire safety inspection registration, emergency rescue organisation designation, personal data protection law governing video and biometric data (a growing compliance burden as facial recognition spreads), defence procurement rules for military work, and building management regulations.

The single most consequential regulatory event for S-1 in the past two years was not a security regulation at all. In January 2026, Korea's Supreme Court ruled in a case brought by fifteen Samsung Electronics retirees that performance incentives must be included in "average wage" for the purpose of calculating severance (Atlas Law). S-1, as a Samsung affiliate with 7,025 employees averaging 11.1 years of tenure, took a one-time retirement benefit charge in Q1 2026 that cut operating profit by 62.4% year on year in that quarter.

Second: the revised Korean Commercial Act, amended in July 2025, which strengthened directors' fiduciary duty to run to all shareholders rather than the controlling shareholder, and limited voting power in director elections. Flashlight Capital has stated explicitly that this amendment is what made its offer possible.

Cyclicality

The subscription base is close to acyclical - a small business that has an alarm keeps paying for it through a downturn, because cancelling is a decision to accept risk rather than to save meaningfully. What is cyclical is new customer formation (fewer new shops opening in a recession), building management tender volume, and above all Security SI and integrated security, which track corporate capital expenditure. When Samsung affiliates build fabs, S-1 gets guard contracts and integration projects. When they pause, it does not. Public sector work is counter-cyclical if anything, since defence and infrastructure budgets are set independently of the business cycle.

Tailwinds and headwinds

Tailwinds: unstaffed premises proliferating; ageing buildings needing monitoring and management; AI making video analytics genuinely useful; access control and biometrics growing at mid-teens rates; anti-drone emerging as a new budget line at airports, ports and nuclear facilities; policy-driven demand from border surveillance and school safety.

Headwinds: minimum wage and labour cost inflation in a headcount-heavy industry; the Supreme Court severance ruling permanently raising retirement liabilities across the Korean corporate sector; CCTV hardware growth decelerating to low single digits; commoditisation of the technology layer as cloud camera vendors sell adequate solutions without a service organisation; and Korean demographics slowly shrinking the domestic customer base.

7. Growth Triggers

Source discipline note: S-1 holds no earnings conference calls. Every item below is sourced to a dated quarterly results release, an accompanying management press statement, a shareholder meeting, or a disclosed corporate filing. Where the source is a sell-side note rather than management, it is labelled as such. Nothing here is inferred.

• Semiconductor and data centre construction projects expected to generate Security SI and integrated security work in H2 2026. Flagged following the Q2 2026 results release as the specific driver expected to reverse the 25.9% year-on-year Security SI decline recorded in the quarter. (Q2 2026 results, released 31 July 2026; Heungkuk Securities note via Asiae, 4 August 2026) - analyst-sourced expectation, not a management guarantee.

• Ministry of National Defence border surveillance system bidding expected in H2 2026. The broader GOP scientific border surveillance improvement programme runs to 2027 with a total budget of 465.1bn won; S-1 is one of three shortlisted bidders alongside SK Telecom and KT. S-1 has already delivered the AI-video-analysis performance upgrade fielded at the 22nd Division GOP and coastal units in December 2024. (Flagged post-Q2 2026, released 31 July 2026; programme detail from BusinessPost and SAT Economy)

• Continued building management contract wins and expanded management scope. This is the single most repeated trigger across the six periods, cited in the Q1 2025, Q2 2025, Q3 2025, Q1 2026 and Q2 2026 releases. In Q1 2026 management specifically added contract renewal price increases as a driver alongside new wins. Repeated in five of six periods.

• Transition of the building management business toward Integrated Facilities Management (IFM). Stated as strategic direction at the March 2026 AGM and restated with the Q2 2026 results. (March 2026 AGM; Q2 2026 release, 31 July 2026) - repeated.

• Expansion of AI and robotics-based products and services, and the shift of security operations "from labour-intensive to technology-centred". The CEO's stated priority on taking office and repeated with the Q2 2026 results.

  "Shifting service delivery from reactive response to predictive prevention using AI capabilities."

  • Jung Hae-rin, March 2026 AGM

  (March 2026 AGM; restated Q2 2026 release, 31 July 2026) - repeated.

• M&A and strategic fund investment to acquire future technologies. Stated at the March 2026 AGM and restated in the 2026 corporate value-up disclosure and with the Q2 2026 results. No target, sector or size has been named. (March 2026 AGM; 2026 value-up plan; Q2 2026 release) - repeated three times, still without specifics.

• Portfolio expansion into fire and safety domains. Identified as a growth vector following the Q2 2026 results. (Q2 2026 results, 31 July 2026, via Heungkuk Securities note)

• Smart School Service scaling. Revenue up 69% year on year through July 2026, deployed with the Gunpo-Uiwang Education Support Office across three Gyeonggi schools plus leak-response consulting for 60 Incheon schools. Driven by increased community evening and weekend use of school gyms and by 23.7% of Korean schools now exceeding 40 years of age. (Company disclosure, 25 August 2026) - the most concrete and best-evidenced growth item on this list.

• Anti-drone solution commercialisation into airports, ports, nuclear plants and military installations. The Ministry of Land, Infrastructure and Transport has planned deployment at major airports and ports by 2026; the Defence Ministry targets military installations. (Product launch, November 2025)

• Samsung AI Door Cam driving home security subscription attachment. Launched December 2025 with Samsung Electronics; 24-hour emergency dispatch sold as a separate paid service on top of the device. (Product launch, 22 December 2025)

• AI-based building energy management systems (BEMS) as a building management upsell. Named at the March 2026 AGM. (March 2026 AGM)

• Overseas market expansion. Named as an expected priority of the incoming CEO at the time of his appointment. No market, timeline or target has been disclosed. (TheBell, 21 November 2025) - the vaguest item on this list, and it has stayed vague for nine months.

8. Key Risks

1. The ownership structure is unstable, and the outcome is genuinely uncertain

Mechanism. S-1's largest shareholder is SECOM of Japan at 25.65%. Five Samsung affiliates hold a combined 20.57% (Samsung SDI 11.03%, Samsung Life 5.34%, Samsung Card 1.91%, Samsung Securities 1.32%, Samsung Fire & Marine 0.97%). Samsung does not control the register but does control the management. On 26 August 2026, Flashlight Capital Partners - a Singapore-based activist fund holding roughly 2% - formally offered to buy the entire Samsung block at 116,000 won per share, a 45% premium to the prior close and above S-1's previous all-time closing high, for total consideration of 906.6bn won. FCP set a deadline of 23 September 2026 (Hedgeweek; Businesskorea).

FCP's argument, per its founder Sanghyun Lee as quoted by Hedgeweek: "Samsung SDI is focused on the capital-intensive battery sector, while the four Samsung financial companies are subject to regulatory restrictions governing cross-shareholdings between financial and non-financial businesses." An FCP representative put it more bluntly to Businesskorea: S-1 is "a representative non-core, low-profit asset that has no rationale for Samsung to continue holding."

Why it is more than noise. The offer is explicitly constructed as a test of the July 2025 Commercial Act amendment, which imposes a duty of loyalty on directors toward all shareholders. Directors of the five Samsung affiliates now have to explain, on the record, why refusing a 45% premium serves their own shareholders. That is a materially different legal position from a year ago.

Calibration. High probability of a disruptive process; genuinely uncertain outcome. Experts cited by Businesskorea expect Samsung to refuse, citing security infrastructure concerns and intra-group governance complexity. But even a refusal has consequences: it forces disclosure of the reasoning, invites follow-on pressure, and puts a public price on the asset. The scenarios here are wide - continuation with an activated activist on the register, a negotiated capital return, a sale of the Samsung block, or a strategic reconfiguration involving SECOM. Every one of them changes the business's relationship with its largest customer group.

2. Losing the Samsung relationship would be worse than losing the Samsung shareholding

Mechanism. These are two separate things and the market frequently conflates them. Samsung affiliates are S-1's shareholders and among its most important customers - integrated security at fabs, building management, the unified badge system, and a call centre segment that is 100% internal. If the shareholding is sold and the commercial relationship weakens over subsequent years, S-1 loses revenue that would be extraordinarily hard to replace. Note also the reverse asymmetry: FCP's own thesis is that S-1 is a "non-core, low-profit" asset to Samsung, which is precisely an argument for the group caring less about the relationship over time.

Calibration. Low probability of an abrupt break (these contracts are embedded and switching a fab's guard force is not a casual decision); moderate probability of slow erosion; high impact if it happens.

3. Labour cost structure, and a permanent step-up in retirement liability

Mechanism. S-1 employs 7,025 people at 11.1 years average tenure. Security guarding and facilities management are headcount businesses where wage inflation flows almost directly to cost of sales, and where a long-tenured workforce means retirement benefit obligations are large and growing. The January 2026 Supreme Court ruling requiring performance incentives to be included in average wage for severance calculation hit S-1 in Q1 2026 with a one-time charge that cut operating profit 62.4% year on year on essentially flat revenue.

Management framed it as a one-off:

"The operating profit decrease reflects temporary impact from one-time cost factors... the security division's steady growth and building management contract expansion show the underlying business fundamentals remain sound."

• S-1 representative, Q1 2026 results, 30 April 2026

They were right about the catch-up charge - Q2 2026 operating profit recovered and grew 4.5% year on year. But the ongoing accrual rate for severance is permanently higher now, and that is not a one-off. This is the highest-probability, moderate-magnitude risk on the list: it will not break the company, and it will keep quietly taxing it.

4. Revenue growth is anaemic and the growth strategy is unspecific

Mechanism. FY2025 revenue grew 3.0%. Q1 2026 revenue was down 0.1% year on year. Q2 2026 revenue was down 1.3%. In a physical security market forecast to grow 12.2% in 2026, that is share loss or mix drag, or both. The stated growth plan - AI, robots, IFM, M&A, overseas, fire and safety - has been repeated across the March 2026 AGM, the value-up disclosure and two quarterly releases without a single named target, market, timeline or acquisition candidate.

FCP made this criticism directly, citing declining market share, weakened profitability, and a lack of clear results in emerging areas including silver care, drones and cybersecurity (Newspim). Management's public counter was that the company had achieved record revenue and operating profit and would continue to pursue "the best methods to raise corporate value" - a response that engages with the scoreboard but not with the strategy question.

Calibration. High probability, moderate drag. This is the base rate of the business, not a shock.

5. Project revenue volatility obscures the underlying business

Mechanism. Security SI is tendered, milestone-recognised and lumpy. It fell 25.9% year on year in Q2 2026 on project timing, which was enough to turn a growing subscription base into a declining reported revenue line. The reverse will happen when the semiconductor, data centre and defence projects land. The risk is not the volatility itself but that it makes the business genuinely hard to read from the outside - which is more damaging than usual for a company that holds no earnings calls and therefore offers no forum to explain it.

6. Technology commoditisation in the mass-market segment

Mechanism. S-1's moat is the dispatch network, not the algorithms. But an increasing share of what a small business actually wants - a camera, cloud recording, phone alerts, remote voice warning - can now be bought from a consumer hardware vendor for a fraction of a monitoring subscription. The unmanned store market that is currently a tailwind is also the segment most exposed to this: an unmanned store owner is by definition comfortable with a technology-only solution. KT Telecop's rise in residential (38.3% versus S-1's 49.0%) is the visible early evidence.

Calibration. Moderate probability, slow-acting, but structurally corrosive because it attacks price per subscriber rather than subscriber count, which makes it hard to see until the mix has already shifted.

7. Capital allocation: a large cash balance doing nothing

Mechanism. FCP's central financial criticism is that S-1 holds cash equivalent to roughly half its market capitalisation while running a dividend policy the fund characterises as inadequate against that balance. The company's own 2026 value-up disclosure commits only to maintaining a 50-60% payout ratio, with no buyback, no cancellation and no ROE target (DigitalToday). The mechanism by which this hurts is not operational - it is that idle capital drags returns on equity, and an under-earning balance sheet in a low-growth business is exactly what invites the kind of campaign now underway.

Calibration. Certain, ongoing, and now the subject of external pressure that may force resolution one way or the other.

8. Concentration in a single, demographically shrinking country

Mechanism. S-1's overseas presence - a Chinese subsidiary, a Mongolian subsidiary, and branches in the UAE, Singapore, Australia and Iraq - was built to follow Samsung construction affiliates, not to win independent market share. Essentially all revenue is Korean. Korea's population is declining and its small business formation rate is under pressure. There is no meaningful geographic diversification and, nine months after "overseas expansion" was named as a CEO priority, no announced plan to build one.

9. Walk the Talk

The six reporting periods used, all confirmed against S-1's own IR earnings archive:

1. Q2 2026 - released 31 July 2026
2. Q1 2026 - released 30 April 2026
3. FY2025 / Q4 2025 - released 29 January 2026
4. Q3 2025 - released 31 October 2025
5. Q2 2025 - released 25 July 2025
6. Q1 2025 - released 25 April 2025

The unavoidable caveat, stated plainly. S-1 does not hold earnings conference calls. There is no forum in which management makes forward-looking commitments, takes analyst questions, or can be held to a number it gave three months earlier. This section therefore assesses credibility against a much thinner evidentiary record than would be normal, and the thinness is itself the most important finding. A company that never puts its CEO in front of investors to answer an unscripted question has, by construction, made itself very difficult to hold accountable. That FCP had to include "CEO-led quarterly earnings calls" among five governance demands in June 2026 tells you this is not an oversight but a settled practice.

The record, period by period

Q1 2025 (25 April 2025) opened the window on a strong note. Operating profit rose 12.8% year on year on revenue up 1.7%. Management attributed the result to large-building security and steady physical security customer growth. No forward guidance was issued.

Q2 2025 (25 July 2025) was the strongest quarter of the six. Operating profit rose 17.6% and net income 20.2%, on revenue up 5.3%. Management's attribution was unusually specific and broad-based: growth in physical security customers in the Security division, and in Infrastructure, new building management contract wins, expanded management scope, and increased Security SI orders - explicitly, growth in every business (Ajunews; BusinessPost). This is the reference point against which everything after should be read.

Q3 2025 (31 October 2025) decelerated. Operating profit growth fell to 3.1% and revenue growth to 2.3%, with net income down 11.1%. Security services and real estate/facility management both grew. Nothing broke, but the growth rate roughly halved in a quarter, and the release offered no explanation for the deceleration and no forward commentary.

FY2025 / Q4 2025 (29 January 2026) closed the year well. Full-year revenue rose 3.0% and operating profit rose 12.1% - the profit growth well ahead of revenue, which is the pattern management had been signalling all year. Q4 specifically was strong on profit, up 16.6%. Alongside the results, the board declared a dividend of 3,200 won per share, up from 2,700 won, a total of 108.19bn won and an 18.5% increase on the prior year (Sidae). Separately, the company filed its 2026 corporate value-up plan, committing to a 50-60% dividend payout ratio.

Judged on the year alone, this is a promise kept. The company said it would return capital through dividends at a high payout ratio, and it did, and it raised the dividend meaningfully.

Q1 2026 (30 April 2026) was the stress test. Revenue was flat (down 0.1%) and operating profit fell 62.4% on the Supreme Court average-wage severance charge. Management's framing:

"The operating profit decrease reflects temporary impact from one-time cost factors... the security division's steady growth and building management contract expansion show the underlying business fundamentals remain sound."

Q2 2026 (31 July 2026) substantially vindicated that claim. Operating profit rose 4.5% year on year and margins improved on both a gross and operating basis, with an independent analyst confirming the quarter represented a full recovery from the Q1 charge and that operating margin ran above the company's normalised level (Heungkuk Securities via Asiae, 4 August 2026). Management had said the charge was temporary. It was.

But the same quarter exposed the weaker half of the record. Revenue fell 1.3% year on year. Security SI fell 25.9%. The Infrastructure division - the growth engine of the previous year - declined 3.3%, with battery factory project revenue delayed. Set against Q2 2025, when management had described growth "across all business divisions," the deterioration in one year is real, and it went unexplained beyond the mechanical attribution to project timing.

Promise versus outcome

The assessment

S-1's management is accurate on the near term and evasive on the medium term, and the distinction is sharp enough to be useful.

On things they can measure and control within a quarter, they are reliable. They said the Q1 2026 charge was a one-off and it was. They said they would maintain a high payout ratio and they raised the dividend. They said the security subscription base was growing and it has grown in every one of the six periods. Nothing in this record suggests management misrepresents results or spins bad quarters into good ones - the Q1 2026 release stated a 62.4% profit decline plainly and explained the cause accurately.

On strategy, the record is much weaker, and the weakness is a pattern rather than a lapse. "AI and robotics", "IFM transition", "M&A and strategic fund investment", "overseas expansion", "fire and safety portfolio expansion" - five growth vectors, all announced, none with a named target, a size, a market or a date attached, several repeated verbatim across multiple disclosures. Nine months after a new CEO named overseas expansion as a priority, there is nothing to evaluate. Revenue has now declined year on year in two consecutive quarters in a market forecast to grow 12.2%, and no explanation beyond project timing has been offered.

The structural point matters more than any individual item. Management has never had to defend a forecast in public, because it does not give forecasts and does not take questions. That is the cleanest possible explanation for why the near-term record looks good and the strategic record is unassessable: they have only ever committed to things they had already banked. A company can be honest and still be unaccountable, and S-1 is currently both. Whether that changes is now substantially in the hands of a Singapore fund with a September deadline.

10. Shareholder Friendliness Index

Dividends. S-1 paid 2,700 won per share for FY2023, held it flat at 2,700 won for FY2024, and raised it to 3,200 won for FY2025, declared on 29 January 2026 for a total payout of 108.19bn won - an 18.5% increase on the prior year's total (S-1 dividend disclosure via KIND, 29 Jan 2026; Sidae; DigitalToday). Two years flat then a real raise, funded by a year in which operating profit grew 12.1%. The company's 2026 corporate value-up filing commits to a 50-60% payout ratio for FY2026, which is genuinely high by Korean standards and qualifies S-1 as a high-dividend enterprise under Korean tax law. The payout ratio does not reveal anything the trend does not, except this: paying out 50-60% of earnings while, on FCP's account, sitting on cash equal to roughly half the company's market capitalisation means the dividend is generous relative to profits and inadequate relative to the balance sheet. Both statements are true simultaneously, and that tension is the entire shareholder-return debate at this company.

Buybacks and dilution. Over the trailing ~90 days (since 30 May 2026), MoatMap's disclosure database records zero buyback transactions. Extending the search externally across the full three-year period, I found no announced or executed share repurchase programme by S-1 in FY2023, FY2024 or FY2025, and none year-to-date in 2026. The FY2025 value-up disclosure contains no buyback or cancellation commitment at all. What does exist is a report that the company is considering buybacks and cancellation in response to market and regulatory pressure on shareholder returns (Bloter, 2026) - consideration, not action. On dilution: shares issued stand at 38,000,000, unchanged, confirmed by the arithmetic on the largest-shareholder disclosure of 9,750,883 shares equalling 25.66%. The FY2025 dividend of 3,200 won totalling 108.19bn won implies roughly 33.8 million dividend-bearing shares, meaning the company itself holds approximately 4.2 million shares (around 11%) in treasury - shares that have been bought at some point historically but never cancelled, and that therefore sit on the balance sheet doing nothing for per-share value. Net change in share count over three years: flat. No dilution, no retirement.

Verdict: Neutral. S-1 pays a real and rising dividend at a high payout ratio, but it has retired no stock in three years, sits on roughly 11% of itself in uncancelled treasury shares and a cash pile an activist puts at half its market value, and has committed to nothing beyond maintaining the payout - which is why it is now facing a public campaign rather than being credited for the dividend.

11. Insider Activities

Source. South Korea's DART disclosure system is API-gated and returns blocked or empty responses to general web search. All transactions below are drawn from MoatMap's cross-market disclosure database (market: KR), current as of 28 August 2026 12:15 UTC, which scrapes the DART "임원·주요주주 특정증권등 소유상황보고서" (Officer and Major Shareholder Securities Holdings Report) and 5% Rule filings directly. Where a transaction has been independently corroborated in the Korean financial press, I have cited that as well. Note that DART filings for officer purchases are typically dated on the filing date rather than the trade date; where press reporting gives the underlying trade dates, I have used both.

Recent transactions (last 12 months, most recent first)

(All rows: DART Officer and Major Shareholder Securities Holdings Report / 5% Rule filings, dates as shown. The Jung Hae-rin trade dates and 장내매수 open-market classification are independently corroborated by DigitalToday, 11 May 2026.)

Buys - reading the signal

Two open-market purchases in the window, both by serving executives, both funded personally.

Jung Hae-rin, President and CEO, 3,500 shares across 7-8 May 2026, approximately 258 million won. This is the transaction that matters, and it needs to be read with care in both directions.

What makes it meaningful: Jung had no prior holding at all. He took office as representative director in March 2026 and bought his first shares eight weeks later, in the open market, across two consecutive trading days, using his own money. There was no vesting event, no option exercise, no grant. A new CEO who buys stock unprompted in his first quarter is making a statement he cannot easily walk back, and the Korean press treated it as one. A first-ever open-market purchase by an incoming CEO with no prior position is a bullish signal.

What tempers it: the absolute size. 258 million won is a modest sum against the compensation of a Samsung president-level executive - realistically a few weeks to a couple of months of total pay, not a life-changing commitment. It is a credible gesture, not a bet-the-house conviction trade. Calling it anything more would be overreading.

Lee Min-jung, Execut ive Vice President, 666 shares on 10 July 2026, approximately 49 million won. A small purchase, but note the shape of it: an odd, non-round lot bought two months after the CEO's, by a serving executive vice president. Odd lots of this kind are usually the residue of someone buying with whatever cash was actually available rather than executing a planned block. It is a genuine open-market purchase, not a housekeeping transaction, and it does not appear to be tied to any vesting or plan event.

Is this cluster buying? Marginally. Two executives buying within roughly two months of each other, both for the first time on the record, in a company where insider purchases are rare, is a mild positive pattern. It is not the classic cluster signal - that would require several board members and the CFO buying in the same window, at meaningful size. Two purchases totalling around 306 million won across a company of 38 million shares is a directional indicator, not a loud one.

Sells - working out the why

Fidelity Management and Research Company LLC, 135,869 shares sold, 5 March 2026, approximately 11.3bn won. This is the only material disposal in the window and it requires the least speculation of any of the transactions here. Fidelity is not an insider in the ordinary sense; it is an external asset manager that has crossed the 5% disclosure threshold and is therefore captured by Korea's 5% Rule reporting regime alongside directors and officers. Its trades reflect portfolio construction - fund flows, position sizing, index weightings, redemptions - not a view formed from inside the company. Reason not disclosed, and in this case no reason is required by the filing: an abbreviated (약식) 5% rule report carries no obligation to state a rationale.

Two contextual points argue against reading it as a negative signal on the business. First, the sale was 135,869 shares against a position that stood at 1,944,996 shares five months later - roughly 7% of the holding trimmed, with 93% retained. Second, and more telling, Fidelity's August 2026 report shows the position at 5.75% of shares outstanding, which is higher than the threshold that triggered disclosure in the first place. Between March and August, Fidelity was not exiting. It was, on the balance of the two filings, adding.

There were no sales by any director or officer of S-1 in the trailing twelve months.

The Fidelity holdings report of 18 August 2026

Listed as "Other" in the transaction feed because it is a position disclosure rather than a trade, this is arguably the most informative single line in the table. A 5.75% stake, filed ten days before this report and eight days before Flashlight Capital's public offer for the Samsung block, means a large, sophisticated, long-only institution is sitting on the register with a position materially above the disclosure threshold while a governance contest plays out. Fidelity's vote, and its posture, is now a live variable in whatever happens by 23 September.

Net assessment

Insiders are net buyers on any measure that counts directors and officers, and the only disposal came from an external institutional holder that appears to have grown its position over the same window.

The activity is narrow: three distinct parties, two of them company executives, one of them an asset manager. It is not broad-based board buying. But the composition is favourable in a specific way that matters. In the twelve months covered, not a single director or officer of S-1 sold a share, and the two who transacted both bought, in the open market, for the first time on record. The CEO's purchase is the standout, and the fact that it came eight weeks into his tenure - before the FCP campaign went public in June, and nearly four months before the August offer - means it cannot be dismissed as a defensive response to activist pressure. He bought when there was nothing to defend against.

What holds the signal back from being strong is size. Roughly 306 million won of combined executive buying is a real gesture from two individuals but a small number against a company of this scale, and there is no CFO purchase, no non-executive director purchase, and no repeat buying by either executive since July.

Read: mildly bullish. A first-time CEO buy with no prior holding, zero insider selling across twelve months, and a 5%-plus institutional holder that trimmed once in March and ended the period larger than it started. Nothing here suggests anyone with information is heading for the exit. Equally, nothing here is the kind of size that tells you an insider believes the stock is dramatically mispriced.

12. Scenarios

Bull case

The governance fight resolves productively rather than acrimoniously. Whether or not Samsung sells its block by 23 September, the board concludes that it cannot keep defending an eleven-percent uncancelled treasury holding and a cash balance that dwarfs the company's investment needs. It cancels the treasury shares, announces a repurchase programme with a real number attached, and pairs it with the first medium-term plan S-1 has ever published - the three-year and five-year frameworks FCP demanded, complete with named target markets. Crucially, it starts holding quarterly earnings calls with the CEO on the line, which does something no press release can: it forces management to commit to things in advance and be measured against them.

Underneath that, the operating story finally converts. The semiconductor and data centre construction cycle lands in H2 2026 and Security SI swings from a 25.9% decline to a growth contributor, taking reported revenue with it. S-1 wins a share of the GOP border surveillance programme running to 2027, which is not merely a contract but the reference that turns anti-drone from a product launch into a pipeline at airports, ports and nuclear facilities as ministry deployment timelines land. Smart School Service keeps compounding off its 69% growth rate as ageing school infrastructure and community out-of-hours gym use both push in the same direction, and the model proves replicable across other public-sector building categories.

The technology transition delivers where it should. AI already filters 78% of control centre signals; pushing that further, plus AI call bots on the customer side, lets S-1 grow subscriber count without growing controller headcount, which is the only way this business escapes the wage inflation and severance accrual that have been quietly taxing it. The Samsung AI Door Cam becomes a genuine subscriber acquisition channel rather than a hardware line, letting S-1 defend residential share against KT Telecop's telecom bundle with a distribution advantage of its own. Blue Asset converts more customers from single-service facility management to full IFM, which raises revenue per building and makes each one materially harder to lose.

By 2029 the company looks structurally different: a subscription base still around half the Korean market but served with far less labour per subscriber, a building management business that has moved up the value chain, a defence and public infrastructure franchise with recurring programme revenue, and a balance sheet that is no longer an argument against the company.

Base case

Nothing much changes, and that is the point.

Samsung declines FCP's offer, citing security infrastructure sensitivity and the complexity of the intra-group structure. FCP stays on the register at around 2%, keeps writing letters, possibly nominates directors, and the campaign becomes a persistent background pressure rather than a decisive event. The board responds with the minimum credible concession - most likely a modest buyback or a dividend increase at the top of the 50-60% payout range - which relieves the immediate pressure without resolving the underlying question of what the cash is for. Quarterly earnings calls either do not happen or happen in a stripped-down form.

Operationally, the subscription base keeps growing at low single digits, because it always has and because nothing in Korean physical security moves quickly. Building management keeps winning contracts and taking renewal price increases, and keeps being the growth line. Security SI keeps whipsawing the reported revenue number quarter to quarter, producing periodic headlines about revenue declining that have very little to do with the business. Some of the H2 2026 semiconductor and data centre work lands, some of it slips into 2027. S-1 wins a portion of the defence programme, not all of it.

The strategic initiatives progress in the way strategic initiatives progress at a company that has repeated them three times without a target: anti-drone books a handful of reference deployments, Smart School grows fast off a small base, the AI Door Cam sells respectably, and "M&A and strategic fund investment" remains a sentence in a disclosure. Overseas expansion continues to mean following Samsung construction affiliates.

Costs stay the central operating challenge. The Q1 2026 severance charge does not repeat, but the higher ongoing accrual does, and wage inflation keeps flowing through a 7,025-person cost base. AI efficiency gains and IFM mix improvement roughly offset it. Margins hold around the levels management has described as normalised, revenue grows slower than the market, and the company keeps paying a good dividend out of a very large cash pile. In three years it is recognisably the same company, slightly more automated, with the same unanswered question about capital.

Bear case

The governance fight goes badly rather than nowhere. Samsung, having been publicly told by an activist that S-1 is "a representative non-core, low-profit asset," internalises the diagnosis without accepting the offer. The affiliates begin a slow, unannounced reduction of their commercial dependence on S-1 - re-tendering building management at some sites, bringing security functions in-house at others, questioning why a captive call centre segment sits inside a listed affiliate at all. None of it is announced. It shows up two years later as integrated security and call centre revenue that stopped growing and then started shrinking, in a business where those contracts were the least contestable revenue S-1 had. The relationship that made S-1 defensible for forty-five years erodes precisely because a third party pointed out that it did not have to exist.

Meanwhile the commoditisation risk arrives faster than expected at the bottom of the market. Cloud camera vendors and telecom bundles keep taking the price-sensitive tier - KT Telecop is already at 38.3% of household security against S-1's 49.0% - and the unmanned store segment, currently the most-cited tailwind, turns out to be the most exposed, because an owner comfortable running a shop without staff is comfortable running it without a monitoring subscription too. S-1's subscriber count holds while revenue per subscriber quietly declines, which is the hardest kind of erosion to see and the hardest to reverse.

The cost structure does not cooperate. Wage inflation continues, the elevated severance accrual is permanent, and the AI efficiency gains that were supposed to offset them turn out to reduce the need for controllers - a few hundred people - rather than for responders and guards, who are the actual headcount. The technology transformation improves the control centre and leaves the cost base largely intact.

Project revenue provides no rescue. S-1 loses the larger tranches of the defence border surveillance programme to SK Telecom or KT, whose network engineering scale wins the systems integration argument even where S-1 wins the video analytics one. The semiconductor and data centre construction work is delayed rather than cancelled, then delayed again. Anti-drone, silver care and cybersecurity continue to produce announcements rather than revenue, exactly as FCP alleged in June 2026.

By the end of it the company still throws off cash, still pays a dividend, still holds around half the Korean market, and still has no answer to the question of what it is going to be. The cash pile grows, the treasury shares stay uncancelled, and the activist - or the next one - comes back with the same letter.