Samsung SDS: The Engine of the Chaebol

I. Introduction & Episode Roadmap

There is a building in 동탄 Dongtan, an hour south of Seoul, where the lights never go out and the air is engineered to a tolerance tighter than most operating theatres. It does not have a Samsung logo on the roof visible from the highway. It does not appear in any Galaxy advertisement. But if the machines inside it stopped for eight hours, the world's largest memory semiconductor operation would begin to lose wafers — silicon discs worth more per gram than most precious metals — because the manufacturing execution systems that tell each machine what to do next would go dark.

That building belongs to 삼성SDS Samsung SDS. And the story of Samsung SDS is the story of what happens when a company is built to serve one customer so completely that it becomes indispensable to that customer, and then spends four decades trying to convince the outside world it can be indispensable to anyone else.

Founded on May 1, 1985 as Samsung Data Systems, the company began as a back-office utility: a place to put the punch-card operators, the mainframe schedulers, and the systems analysts scattered across a sprawling Korean conglomerate.¹ Forty-one years later it employs 26,157 people, operates from 56 offices across 40 countries, and reported consolidated revenue of KRW 13.93 trillion in 2025 — roughly $9.5 billion — with operating profit of KRW 957.1 billion.¹² Its market capitalisation as of August 2026 sits around KRW 17.1 trillion, or a little under $12 billion.³

Those numbers conceal something important, and it is the first thing any investor has to internalise about this company: Samsung SDS is not one business. It is two businesses with almost nothing in common, stapled together in a single P&L.

The first is IT Services — cloud infrastructure, managed services, enterprise applications, and now generative AI platforms. In 2025 it produced KRW 6.54 trillion of revenue, a little under half the total, and the overwhelming majority of the profit.² The second is Logistics BPO, a global freight-forwarding and contract-logistics operation run through the digital platform 첼로스퀘어 Cello Square. It produced KRW 7.39 trillion of revenue — more than half the top line — and just KRW 130 billion of operating profit.² Roughly 53% of the revenue delivers roughly 14% of the earnings.

That structural mismatch is why Samsung SDS trades the way it does. A screener sees a KRW 14 trillion revenue company with a 6.9% operating margin and files it under "Korean industrial conglomerate subsidiary." An analyst who strips out logistics sees a software and cloud business with double-digit margins, net cash, and a captive customer that happens to be one of the most capital-intensive manufacturers on earth. Both are looking at the same company.

Layered on top is the governance question that has shadowed the stock since before it was even listed. 삼성SDS Samsung SDS sits at the intersection of three things the Lee family cares enormously about: the operational security of 삼성그룹 Samsung Group, the manufacturing execution of 삼성전자 Samsung Electronics, and — for most of the last decade — the family's own liquidity needs. In a 재벌 chaebol, a captive IT subsidiary is never purely an operating asset. It is also a balance-sheet instrument.

This piece traces the arc: from building Samsung's digital nervous system in the 1980s and 1990s, to the logistics pivot that quadrupled revenue and halved margins, to the 2014 IPO that made SDS briefly Korea's sixth most valuable listed company, to the grinding decade-long transformation from systems integrator to cloud provider, to the 2023 acquisition of 엠로 Emro, and finally to the current moment — an enterprise AI and GPU infrastructure build-out funded partly by KKR, with capacity targets that would make Samsung SDS one of the larger AI compute landlords in Asia.

The question threaded through all of it is simple and unresolved: is Samsung SDS a genuine enterprise technology company that happens to have a chaebol parent, or a chaebol utility that has learned to speak fluent cloud? The evidence points both ways, and the last two years have made the argument more interesting, not less.

It starts, as most chaebol stories do, with a founder deciding that the group needed something it did not yet have.

---

↑ Back to Top

II. Founding Context & The Enterprise Nervous System (1985–2010)

In 1985, South Korea's per-capita GDP was around a tenth of Japan's, and Samsung was a domestic trading house with electronics ambitions rather than a global brand. Inside the group, every affiliate—from the insurance arm to the fledgling semiconductor operation—purchased its own computers, hired its own programmers, and built incompatible internal systems.

이병철 Lee Byung-chull, Samsung's founder, was in the final years of his life and embarking on the largest gamble of his career: pushing Samsung into DRAM manufacturing despite widespread skepticism. Semiconductor manufacturing cannot be managed on paper ledgers; it relies on yield curves, lot tracking, and precise process control, where a single misstep across a batch of wafers can ruin millions of dollars of inventory. Samsung Data Systems was incorporated on May 1, 1985, to consolidate the group's fragmented IT operations and establish a single technical backbone.¹

For its first decade, the work remained unglamorous: managing payroll, general ledgers, and order entries. The company was renamed Samsung SDS in 1997, as Korea's IT sector professionalized and the group began viewing software as a core strategic input rather than an overhead cost.¹

Why the fabs changed everything

The transformation that reshaped Samsung SDS was industrial rather than commercial.

As Samsung Electronics expanded from a single fab into multiple facilities—and from DRAM into NAND flash, contract foundry manufacturing, and display panels—it required sophisticated manufacturing execution systems (MES). An MES acts as the traffic control tower for a semiconductor facility. It tracks every wafer lot, assigns specific tools, runs exact recipes, monitors sensor measurements, and flags machine drifts before defects compound. In a modern semiconductor fab, a wafer undergoes roughly 800 process steps over three months. An error in sequencing results in immediate scrap.

Samsung SDS designed and operated these control systems. Above them, it integrated the enterprise resource planning (ERP) software that converts factory output into inventory records, invoices, and financial statements, while managing the underlying network and security infrastructure.

This architecture created a deeply entrenched competitive lock-in. The primary switching cost is not software licensing or migration expense; it is operational risk. Replacing an MES in an active fab requires running old and new systems in parallel through a full production cycle on a line generating billions of dollars in output, with zero tolerance for lost wafers. Consequently, replacement has rarely, if ever, been seriously contemplated by group procurement.

That operational risk aversion within high-stakes manufacturing represents the core foundation of the Samsung SDS business model—far more than brand, proprietary technology, or scale.

The captive market, and its price

The primary commercial benefit was a guaranteed order book. Samsung Group affiliates provided a floor under revenue that was largely insulated from broader Korean IT spending cycles. Group demand also provided SDS with an exceptionally demanding internal testbed. Software built to handle the scale and precision of a Samsung semiconductor fab had already undergone more rigorous stress-testing than standard commercial enterprise software.

However, heavy reliance on captive demand introduced substantial regulatory and political vulnerabilities. In South Korea, 일감 몰아주기—the practice of funneling internal work to group affiliates, particularly those in which controlling family members hold equity—has drawn continuous scrutiny from the Fair Trade Commission and political reformers. Samsung SDS has operated under this regulatory spotlight throughout its history as a public company. The proportion of standalone revenue derived from group affiliates peaked at roughly 87% in 2017. By 2022, disclosures compiled from Fair Trade Commission data showed intra-group transactions still generated KRW 4.2 trillion—or 81.5% of the parent entity's revenue—with Samsung Electronics alone accounting for KRW 2.5 trillion.⁴

For investors, captive revenue presents a double-edged dynamic. While affiliate contracts ensure steady volume, they are subject to strict transfer-pricing regulations that limit profit margins. As a result, SDS operates an internal monopoly that it cannot price like a commercial monopolist.

The ceiling on the SI model

By the mid-2000s, Samsung SDS faced a structural limitation inherent to traditional systems integration (SI).

Conventional SI functions largely as a labor-arbitrage service model. Providers bid on fixed-price contracts, assign engineering staff, and generate profit by minimizing billable hours relative to the contract value. Margins remain structurally constrained, revenue is non-recurring, and scaling top-line growth requires linear headcount expansion. In South Korea, this market was divided among three similar conglomerates with captive IT arms: 엘지씨엔에스 LG CNS for LG Group, SK㈜ C&C SK Inc. C&C for SK Group, and later 현대오토에버 Hyundai AutoEver for Hyundai Motor Group. Because affiliates rarely awarded contracts outside their parent groups, market share remained static.

By 2010, Samsung SDS possessed a stable captive business but lacked a clear external growth engine. The domestic SI market had matured, affiliate revenue was constrained by regulatory oversight, and the group's flagship manufacturing arm was preparing to introduce an entirely new operational challenge.

---

↑ Back to Top

III. The Logistics Pivot: Inventing Logistics BPO (2011–2014)

By 2010, Samsung Electronics was moving physical goods at a scale that had outgrown its internal systems—shipping tens of millions of mobile phones, televisions manufactured in Mexico and Slovakia, and appliances, components, and semiconductors across global trade routes.

Yet the freight booking process remained fragmented and surprisingly manual for an enterprise that measured semiconductor yields to four decimal places. Regional teams negotiated independently with local forwarders, shipping rates were opaque and tied to the spot market, and cargo visibility often depended on emails chained through intermediaries. When ocean freight rates spiked, Samsung lacked group-wide bargaining power; when rates fell, it had no centralized mechanism to capture the savings quickly.

Samsung SDS framed this operational challenge not as a maritime shipping problem, but as a data management issue.

From software vendor to freight principal

The company developed Cello, a proprietary logistics platform designed to sit above carriers rather than compete with them—orchestrating bookings, tracking shipments, optimizing routes, and managing warehouses across the group's global network. Crucially, Samsung SDS did not merely license the software; it assumed direct operational responsibility as a fourth-party logistics (4PL) provider. Under this model, Samsung SDS managed end-to-end supply chains—spanning international transport, inland freight, warehousing, and last-mile delivery—with Cello functioning as the underlying operating engine.⁵

The strategic rationale was straightforward. As the single booking entity for one of the world's largest shippers, Samsung SDS wielded immense aggregate purchasing power. Owning the data layer provided network visibility that no individual freight forwarder possessed. Furthermore, for a company whose captive IT revenue faced regulatory scrutiny in South Korea, logistics business process outsourcing (BPO) offered a path to expand top-line revenue rapidly without raising software charges for Samsung Electronics.

The financial metamorphosis

This operational shift fundamentally altered the company's financial profile.

Under principal freight-forwarding accounting, revenue is recognized on a gross basis. When Samsung SDS books container shipping, the full freight charge is recorded as top-line revenue, while the vast majority flows out directly to ocean carriers as cost of sales. Samsung SDS retains only the net spread. In freight forwarding, operating margins are thin—typically ranging between 2% and 5%.

Combining this low-margin, high-volume logistics business with high-margin IT services reshaped reported financial metrics. Revenue expanded rapidly, while consolidated operating margins declined—reflecting the mathematical blending of two distinct operating models rather than a deterioration in IT services performance. The growth of logistics volumes also elevated the company's reliance on group business: affiliate transactions rose from 51.0% of revenue in 2011 to 56.4% in 2012, reaching 65.5% by 2013.⁴

Consolidated margins can obscure the underlying economics of Samsung SDS. Blending an IT services business that generates double-digit operating margins with a freight business operating at lower single-digit margins creates an aggregate figure that reflects neither segment accurately. Investors focusing solely on consolidated margins risk underestimating the software business, while those ignoring logistics overlook its working capital requirements, operational demands, and exposure to freight-market cyclicality.

This dynamic became evident during the pandemic freight boom. Consolidated revenue reached KRW 17.23 trillion in 2022 as ocean freight rates surged, yet operating profit reached KRW 916.1 billion—only a modest increase from the KRW 808.1 billion earned in 2021 on revenue of KRW 13.63 trillion.⁶ Roughly KRW 3.6 trillion in incremental revenue yielded minimal incremental profit. When freight rates normalized in 2023, consolidated revenue fell back to KRW 13.28 trillion, accompanied by a corresponding decline in operating profit.⁶ The logistics segment effectively serves as a top-line multiplier with minimal earnings leverage.

Understanding this structural division is critical to evaluating the company's financial evolution, establishing the backdrop for its 2014 initial public offering, which arrived just as the logistics transformation took hold.

↑ Back to Top

IV. The 2014 IPO: Governance, Succession, and the Chaebol Discount

On November 14, 2014, Samsung SDS listed on the Korea Exchange in what was then one of the country's largest initial public offerings. Priced at KRW 190,000 per share—the top of the marketed range—the offering raised roughly $1.1 billion.⁷⁸

Trading momentum took over immediately. The stock more than doubled intraday before closing at KRW 327,500, 72% above the offer price. That gave Samsung SDS a market capitalization of KRW 25.3 trillion, making it South Korea's sixth most valuable listed company on day one.⁹ Brokerages quickly initiated coverage with 12-month price targets as high as KRW 500,000.⁹

Yet nearly twelve years later, in August 2026, the shares traded around KRW 221,500—roughly a third below that first-day close.³ That performance reflects a fundamental gap between initial market expectations and operating realities.

What investors were actually buying

The divergence between market euphoria and subsequent share performance stemmed less from operating metrics than from the transaction's perceived purpose.

Bloomberg's coverage of the debut underscored that market participants were primarily betting on a broader corporate restructuring.⁹ The shareholder register drove that thesis: 이재용 Lee Jae-yong, 이부진 Lee Boo-jin, and 이서현 Lee Seo-hyun—the three children of Chairman 이건희 Lee Kun-hee, who had been incapacitated earlier that year—held substantial personal stakes in Samsung SDS. With South Korean inheritance taxes on controlling stakes effectively exceeding 50%, investors anticipated a liquidity event or structural rearrangement.

The listing served two primary functions for the controlling family. First, it established a public market valuation for family-held shares that could eventually be sold or pledged to satisfy inheritance tax obligations. Second, it created a listed vehicle whose equity could theoretically be integrated into a future group holding structure.

Consequently, the IPO valuation reflected operating fundamentals alongside significant option value tied to a corporate restructuring that was never formally announced nor ultimately executed in the form markets anticipated.

The 2016 split-off fight

This structural tension surfaced eighteen months later. In mid-2016, Samsung SDS confirmed it was evaluating a plan to split off its logistics division.¹⁰ Market analysts anticipated that the IT solutions business would merge toward Samsung Electronics while the logistics arm would transfer to 삼성물산 Samsung C&T, positioning Samsung C&T as the group's de facto holding company.

Minority shareholders vigorously opposed the proposal. The Samsung SDS Small Investors' Group sent delegations to corporate headquarters in June 2016, alleging that executives had previously pledged to avoid physical corporate divisions that harmed shareholder value before reversing course to consider all options, including "the physical division of the firm."¹¹

Their opposition was grounded in financial logic: logistics generated the majority of consolidated revenue, so carving out that division into an entity with higher controlling-family ownership would effectively transfer corporate value away from public SDS shareholders. The corporate division was ultimately shelved.

The dispute highlighted a persistent structural risk: Samsung SDS's corporate perimeter remained subject to broader group ownership priorities rather than pure commercial logic. For public investors, this dynamic reinforced the "Korea discount," reflecting a governance risk premium required to offset potential structural reorganizations.

The capital allocation record

Capital allocation has presented a secondary headwind to valuation. Samsung SDS has consistently maintained a large net cash position. At the end of 2025, the company held KRW 6.38 trillion in cash and short-term investments against KRW 817 billion in total debt, which consisted almost entirely of capitalized lease obligations.⁶ Total equity stood at KRW 10.26 trillion.⁶

Holding cash reserves of that scale on low-yield deposits created a persistent drag on return on equity. In response, management established a three-year capital return policy for 2025 through 2027 targeting a dividend payout ratio of approximately 30%, tied to annual earnings.¹² Under this policy, the 2025 dividend was increased roughly 10% to KRW 3,190 per share.²

A 30% payout ratio for a firm holding more than a third of its equity in cash represents a conservative distribution framework by global enterprise-software standards, particularly given the absence of share repurchases or cancellations. Whether this cash balance functions as excess capital or strategic dry powder has remained a key analytical question for investors as the company enters its next phase of enterprise AI and cloud expansion.

↑ Back to Top

V. The Cloud Transformation: From SI to CSP & MSP (2015–2023)

Between roughly 2012 and 2020, global IT services incumbents faced a fundamental market shift: enterprise clients stopped buying physical servers for on-premises data centers and hiring systems integrators to write bespoke software. Instead, corporations began renting compute capacity on demand from hyperscalers like Amazon, Microsoft, and Google.

For Samsung SDS, this transition threatened the traditional systems integration model. While SI monetized operational complexity through billable engineering hours on custom deployments, cloud computing prioritized standardization, elasticity, and self-service. Corporate IT budgets did not vanish, but spending migrated from custom application development to consumable infrastructure.

Samsung SDS responded by pursuing a dual-track cloud strategy, balancing internal security requirements against external commercial expansion.

Playing both sides of the cloud

First, the company established itself as a Cloud Service Provider (CSP) by building Samsung Cloud Platform (SCP) and the proprietary data centers required to run it.¹³ The primary driver was operational security rather than pure economics. Sensitive Samsung Group workloads—including semiconductor fab recipes, yield data, and unreleased product designs—represent critical trade secrets that could not be hosted on public infrastructure operated by third-party vendor platforms. Virtualizing those manufacturing assets required maintaining physical ownership of the underlying hardware within the group's security perimeter.

Second, Samsung SDS expanded as a Managed Service Provider (MSP), helping external enterprise clients migrate to and manage workloads on AWS, Azure, and Google Cloud. Although MSP operates at lower margins and positions Samsung SDS as an intermediary alongside competing public cloud platforms, it provided a channel for external growth.

Samsung SDS has cited market research from Gartner, IDC, and Frost & Sullivan to support its claim of holding the top market share in Korean managed cloud services.¹ Recent MSP growth has extended beyond Samsung affiliates into non-group sectors, including cloud migrations for Korean financial institutions, public-sector generative AI initiatives, and enterprise resource planning and supply chain deployments for industries such as shipbuilding.²¹⁴

Winning competitive tenders in commercial banking and government projects demonstrates that technical capabilities developed within the group translate into the broader domestic market. However, the MSP model carries structural limitations. As an intermediary, Samsung SDS exercises limited pricing power because underlying infrastructure costs are established by global hyperscalers, leaving differentiation centered on service quality and migration expertise within a domestic market of roughly 50 million people.

The capital intensity trap

The CSP strategy introduced a distinct operational risk: capital intensity.

Unlike systems integration, which relies on variable labor costs that scale down during demand contractions, data center development requires large upfront capital commitments years before generating revenue. Facility returns depend entirely on utilization rates, making underused compute capacity a direct drag on return on invested capital.

Despite those capital demands, Samsung SDS built out its physical footprint, headlined by the high-density 동탄 데이터센터 Dongtan Data Center campus in Gyeonggi Province. Following the opening of its first building, the addition of a second facility expanded total campus capacity to 40 megawatts when it entered operation in the first quarter of 2026.¹⁵ Procurement contracts for the Dongtan build totaled hundreds of billions of won.¹⁶

By 2023, the outline of management's strategy was clear: maintain secure private infrastructure for captive group workloads, sell managed services to external enterprise clients, and expand physical data center capacity ahead of future compute cycles. That infrastructure laid the foundation for the sudden demand spike in enterprise generative AI.

Before capitalizing on that compute cycle, however, the company had to address another pressing issue: how to deploy its growing cash reserves.

---

↑ Back to Top

VI. M&A & Capital Deployment: Benchmarking the 엠로 Emro Deal

For most of its listed history, Samsung SDS's approach to acquisitions could be summarized in one word: avoidance.

The company accumulated cash, paid modest dividends, and made minor technology investments. It left billions of dollars sitting on the balance sheet while institutional investors noted, with growing frustration, that earning a low single-digit return on cash while trading at a discount to book-adjusted peers reflected a capital allocation problem rather than an operating issue.

On March 15, 2023, Samsung SDS altered that stance.

The deal

The company acquired a 33.4% stake in 엠로 Emro, Korea's leading vendor of procurement supply chain management software, becoming its largest shareholder with board representation.¹⁷ The consideration was KRW 111.8 billion for approximately 3.74 million shares—around $86 million at the time.[^18] Emro was a focused business with roughly 300 employees, specializing in procurement consulting, deployment, and software development.¹⁷

Against a multi-billion-dollar cash reserves balance, this was a modest deployment—representing under 2% of total cash holdings. The transaction highlighted Samsung SDS's cautious institutional culture, executing its first major software acquisition in the minimal increment required to test the investment thesis.

The strategic logic, which is better than the size suggests

The underlying rationale was operational completeness. Enterprise supply chain management spans three distinct layers, and Samsung SDS previously held capabilities in only two.

Planning answers what to manufacture and where to deploy it through demand forecasting, inventory optimization, and allocation. Samsung SDS addressed this layer partly through its proprietary Nexprime SCM suite and partly through a partnership with American planning software developer o9 Solutions.¹⁸

Procurement governs sourcing, supplier management, and contract compliance—determining vendor selection, pricing, and terms. This was the gap that Emro filled.

Execution manages physical transport and warehousing, operating through the Cello and Cello Square platforms built during the logistics expansion.

Unifying all three layers allowed Samsung SDS to offer an integrated supply chain suite rather than isolated point solutions, shifting sales conversations from domestic IT integration to end-to-end supply chain management. President and CEO 황성우 Hwang Sung-woo framed the strategic goal explicitly at the announcement, stating that combining forces with Emro would expand the integrated SCM platform globally and solidify Samsung SDS's position as a SaaS solution provider.¹⁷

Testing the valuation and the promise

On price, the transaction appeared reasonable rather than aggressive. Samsung SDS acquired effective control of a listed Korean software vendor at an implied total enterprise valuation of roughly $260 million. By contrast, Western supply chain assets commanded premium valuations during the same cycle—highlighted by Thoma Bravo's $8 billion buyout of US procurement vendor Coupa. Samsung SDS secured a domestic market leader at a Korean small-cap valuation multiple.

The strategic question is whether a Korean procurement leader can scale internationally. Procurement software is deeply tied to local commercial practices, supplier networks, tax rules, and regulatory frameworks. Emro's domestic strength—deep integration with Korean industrial manufacturers—does not automatically translate into competitive wins against SAP Ariba or Coupa in Germany or the United States.

Post-acquisition integration has proceeded cautiously. Samsung SDS preserved Emro as a separate listed entity rather than absorbing it directly, focusing on commercial collaboration, including a joint arrangement with o9 Solutions to attack the global SaaS-based SCM market.¹⁸ In June 2026, Samsung SDS appointed Executive Vice President Sangwon Jo as Emro's CEO—signaling a shift from passive strategic oversight to direct operational direction.¹⁹

Three years after the deal, the Emro acquisition stands as a logical portfolio fill-in that has yet to prove itself as a major global SaaS platform. Management has not disclosed standalone international SaaS revenue metrics for the combined offering, indicating that overseas commercial expansion remains in its early stages.

The acquisition demonstrated that Samsung SDS was willing to execute strategic M&A. However, addressing the company's next growth phase—scaling AI compute infrastructure—would require far more capital than the KRW 111.8 billion committed to Emro.

↑ Back to Top

VII. The Enterprise AI Surge: FabriX and Brity Copilot (2024–Present)

In 2023, an incident inside Korean corporate IT circles sharply focused executive attention on data security. Employees at a major manufacturer had reportedly pasted proprietary source code and internal meeting transcripts into a public chatbot for assistance. Once uploaded to an external server, the sensitive data could not be retrieved.

For Samsung Group, data security is an operational priority. Leading-edge semiconductor process specifications—including precise chemical recipes, tool tolerances, and equipment parameters—represent decades of research capital and determine the financial viability of multi-billion-dollar fabrication facilities. Any enterprise system that allows employees to query sensitive data through generative AI introduces potential leakage vectors if not strictly contained.

Yet the potential productivity gains from enterprise generative AI made adoption irresistible. Navigating this tension between operational security and workforce efficiency led Samsung SDS to develop two complementary software platforms.

What FabriX actually is

Behind the marketing framing, FabriX operates as an enterprise data broker. It bridges a company's internal systems—including ERP records, manufacturing data, proprietary documents, and workflow tools—with large language models capable of reasoning over that information.²⁰

A helpful framework is to view FabriX as a secure intermediary within corporate networks. When an employee submits a natural-language query, FabriX identifies the relevant internal databases, retrieves the necessary records, filters sensitive data according to role-based access rules, and routes the query to an approved internal or external model before returning the response. To the user, the interface resembles a standard chatbot; behind the scenes, every interaction is a logged, authenticated query against enterprise systems of record.

The commercial rationale reflects a key market reality: in enterprise AI, underlying language models are increasingly commoditized, whereas value accrues to integration and governance. Frontier models have become widely available at similar price points, but managing proprietary data architecture, access permissions, and systems of record requires deep institutional knowledge. Samsung SDS has developed that data mapping across Samsung Group over four decades.

Reflecting management's strategy of treating underlying models as consumable inputs, Samsung SDS made FabriX available on Microsoft Azure in September 2024, signaling that the platform is built to operate across whichever cloud infrastructure a customer chooses.[^22]

Brity Copilot and the productivity layer

The second platform, Brity Copilot, launched in April 2024, targets everyday corporate collaboration by applying generative AI to draft and summarize emails, transcribe meetings, synthesize minutes, organize documents, and automate routine administrative tasks.²⁰

Initial adoption metrics released by Samsung SDS offer concrete indicators of user engagement. More than 100 companies adopted the software, with active users logging an average of 14 interactions per day.²⁰ According to company disclosures, preparation time for meeting minutes dropped by more than 75%.²⁰

An average of 14 daily interactions per user indicates regular operational integration rather than speculative testing. Samsung SDS also benefited from an internal deployment runway: Samsung Group employs several hundred thousand workers, providing an immediate, large-scale reference deployment before pitching external corporate clients.

However, Samsung SDS faces direct competition from global incumbents like Microsoft and Google, which bundle similar generative features directly into established productivity suites. While Korean-language tuning and local workflow customization offer a competitive advantage domestically, that edge provides limited differentiation in international markets.

From projects to platforms — and the frontier-model alliances

From a financial perspective, the strategic objective of this AI initiative is improving revenue quality. Traditional custom AI consulting resembles legacy systems integration—project-based, labor-intensive, and non-recurring. By contrast, per-seat software licensing and consumption-based compute services deliver recurring, higher-margin revenue.

Executing this model shift required Samsung SDS to secure two foundation layers: frontier model partnerships and compute capacity.

Rather than committing to a single model provider, Samsung SDS established collaboration frameworks with Google Cloud and OpenAI. On July 25, 2026, the company announced that CEO 이준희 Lee Jun-hee had signed a strategic partnership agreement with Anthropic the previous day at an AI summit in San Francisco, making Samsung SDS the first Korean company to enter a formal business partnership with the AI developer.²¹²² The agreement covered joint commercial development in South Korea, technical training for applied AI engineers on the Claude platform, and the distribution of Claude Enterprise across Samsung Group companies—where roughly 70,000 employees across 20 affiliates were already using the service.²²

Deploying 70,000 seats across group affiliates demonstrates that internal AI adoption is generating actual software revenue rather than serving merely as promotional positioning. At the same time, it highlights the firm's ongoing reliance on captive group demand. Captive adoption confirms technical viability, but establishing broad commercial validation will require winning non-affiliated enterprise accounts.

That challenge turns directly on physical infrastructure, capital investment, and capacity scale.

↑ Back to Top

VIII. Current Financial & Segment Deep Dive

The 2025 results, published on January 26, 2026, were the definition of a mixed year that management framed as a good one.²

Consolidated revenue was KRW 13.93 trillion, up 0.7%. Operating profit was KRW 957.1 billion, up 5.0%.² A company growing revenue by less than one percent is not a growth company, and the headline flatters nothing. But the composition underneath was moving in the direction management has promised for a decade.

IT Services revenue rose 2.2% to KRW 6.54 trillion, with cloud revenue up 15.4% to KRW 2.68 trillion — meaning cloud alone was now roughly 41% of the IT Services segment and had become the company's principal growth engine.² Logistics revenue fell 0.5% to KRW 7.39 trillion as maritime freight rates continued to decline, and logistics operating profit fell 6.2% to KRW 130 billion.²

That KRW 130 billion figure deserves a moment. It represents about 13.6% of group operating profit generated by 53% of group revenue, at a segment operating margin of roughly 1.8%. Every won of logistics revenue is contributing under two jeon of operating profit. The segment is not a disaster — it is cash-generative, it is strategically useful, and its digital platform is growing — but as a profit business it is close to immaterial, and as a revenue business it dominates every consolidated ratio an investor might calculate.

The 2026 quarters: one ugly, one encouraging

The first quarter of 2026 looked alarming on the surface. Revenue fell 3.9% to KRW 3.35 trillion and operating profit collapsed 70.8% to KRW 78.3 billion.²³ The explanation was a one-time KRW 112 billion retirement benefit provision following a change in how retirement allowances are calculated.²³ Adding that back gets underlying operating profit to roughly KRW 190 billion — still weak, but a different story entirely. Cloud revenue in the quarter was KRW 690.9 billion, up 5.8% year on year, with CSP up 12% and MSP up only 4%.²³

An accounting-driven earnings collapse is exactly the kind of event where management credibility is tested. Samsung SDS disclosed the driver, quantified it, and told the market that the second quarter would improve on public-sector GPUaaS demand, financial-sector revenue, and government-wide AI service rollout.²³

The second quarter, reported on July 30, 2026, delivered on the direction if not the magnitude. Revenue rose 5.9% to KRW 3.72 trillion; operating profit rose 0.7% to KRW 231.8 billion; net profit rose 4.6% to KRW 184.1 billion.²⁴ IT Services revenue was KRW 1.76 trillion, up 5%, with cloud revenue up 17% to KRW 779.4 billion.²⁴ Within cloud, CSP grew 24% on Samsung Cloud Platform demand and expanded GPU-as-a-Service engagements, while MSP grew 17% on financial-sector AI transformation projects and shipbuilding ERP deployments.²⁵ Logistics revenue rose 6.6% to KRW 1.96 trillion, helped by air forwarding, contract logistics, and external business won through Cello Square.²⁵

Note what the second quarter did not show: operating leverage. Revenue up 5.9%, operating profit up 0.7%. The mix is improving and the margin is not, because the company is simultaneously spending heavily to build AI infrastructure. That is a defensible trade, but it is a trade, and it should be evaluated as one.

The non-captive metric

The number that generated the headlines was this: cloud revenue from customers outside Samsung Group grew 75% year on year in the second quarter of 2026 — the sharpest quarterly acceleration the company has reported in this metric.²⁵²⁴

This is the single most important disclosure Samsung SDS makes, and it deserves careful handling rather than celebration.

What it genuinely shows: external enterprises and public agencies are buying compute and managed services from Samsung SDS in competitive situations. The named drivers are specific and verifiable in kind — an AI agent banking project for 우리은행 Woori Bank, a generative AI system for the 한국수출입은행 Export-Import Bank of Korea, data platform projects across major commercial banks, and generative AI services extended to dozens of government agencies.¹⁴² These are not related-party transactions.

What it does not show: the base. Samsung SDS has not disclosed the absolute won value of external cloud revenue, only the growth rate. A 75% increase on a small base is a promising early signal; the same percentage on a large base would be transformative. Without the denominator, an investor cannot tell which. This is a disclosure gap, and it is a fair thing for a sceptical investor to press management on.

The context that makes the growth plausible rather than suspicious is Korea's AI compute shortage. GPU capacity has been scarce, the government has been actively funding domestic AI infrastructure, and Samsung SDS was first in Korea to deploy NVIDIA's B300 accelerators in March 2026, with full commercial operation of the B300 GPUaaS platform reached in the third quarter.²⁶²⁵ When supply is short, whoever has racked hardware wins business. The durability of that advantage as capacity floods in is a genuinely open question.

The balance sheet, and what changed in April 2026

Samsung SDS entered 2026 with the same fortress balance sheet it has always had — cash and short-term investments of KRW 6.38 trillion, no meaningful borrowings, strong cash conversion.⁶ At the 41st annual general meeting on March 19, 2026, CEO Lee Jun-hee told shareholders the company viewed 2026 as "a pivotal year that will determine leadership in AI and cloud," and laid out plans to deploy the cash into a new data centre in 구미 Gumi, the national AI computing facility, and GPU servers at Dongtan.²⁷ Shareholders also approved raising the convertible bond issuance ceiling from KRW 67 billion to KRW 1.5 trillion — a technical resolution whose purpose became obvious four weeks later.²⁷

On April 15, 2026, Samsung SDS announced it would issue KRW 1.22 trillion — about $820 million — of convertible bonds to KKR, with a conversion price of KRW 180,000 per share, an 18.8% premium to the prior close, funded principally from KKR's Asia Fund IV and expected to close in the second quarter.²⁸ KKR also agreed to a six-year active advisory role covering M&A, capital allocation, and global growth strategy.²⁸ The shares rose more than 20%.²⁹

That market reaction is worth decoding, because a convertible bond is dilutive and a 20% pop is not the normal response to dilution. Investors were not pricing the money. Samsung SDS did not need money; it had KRW 6.4 trillion of it. They were pricing three other things: an outside institutional investor with global technology franchises validating the AI strategy after diligence, a six-year advisory relationship that inserts an external voice into the capital allocation process of a company long criticised for having none, and — most tellingly — a signal that management intended to spend aggressively rather than hoard.

For a stock whose discount has been substantially a governance discount, the arrival of a credentialed outside partner in the capital structure was the news. Whether it changes behaviour is a 2027–2028 question.

---

↑ Back to Top

IX. Competitive Strategy & Moat Analysis

Strip Samsung SDS down to its structural foundations and ask the central question for any long-term investor: what, specifically, prevents a competitor from replicating this business?

Hamilton Helmer's 7 Powers, applied honestly

Switching costs — genuine and dominant, but narrow. The MES and ERP entanglement inside Samsung Electronics' fabs represents as formidable a switching cost as exists in enterprise technology, for reasons already established: the cost of migration is measured in operational risk to multi-billion-dollar production lines, not software licensing fees. However, this advantage applies to a single customer group. It does not transfer externally. It provides Samsung SDS with an unassailable position in roughly a third of its business, but offers no competitive edge in the remaining two-thirds it aims to grow.

Cornered resource — real, but shrinking in relative importance. Samsung SDS maintains direct access to manufacturing telemetry, security protocols, and operational data flows that third-party cloud providers are structurally prohibited from inspecting. No external vendor is permitted inside a proprietary semiconductor process recipe. That creates an information asset competitors cannot acquire at any price. Yet its commercial value remains constrained by intra-group transfer-pricing rules and by the fact that it yields internal expertise rather than a licensable product. Samsung SDS knows how to manage a semiconductor fab's IT infrastructure; it cannot commercialize that proprietary domain knowledge with competing chipmakers like TSMC.

Scale economies — moderate, and improving. Spreading data center capital expenditure and software R&D across Samsung Group's massive internal footprint grants Samsung SDS a lower unit cost of compute capacity than domestic peers. Against global hyperscalers operating gigawatt-scale infrastructure, however, it holds no cost advantage. The company's AI infrastructure roadmap — expanding from 110 megawatts of capacity today to 230 megawatts by 2029 and over 800 megawatts by 2031 — represents an attempt to convert a domestic scale advantage into a regionally relevant footprint.³⁰

Process power — plausible, but largely unproven externally. The routing and logistics optimization logic embedded within Cello Square reflects over a decade of accumulated operational refinement. The platform reached 24,625 subscribing companies at the end of 2025, up 27% year-on-year, while adding analytical AI tools for predictive departure and arrival timing alongside early transshipment bottleneck detection.²³¹ Sustained subscriber growth on an external platform offers tangible evidence of standalone software value. What remains undisclosed, however, is the gross margin or take-rate realized on that external volume — leaving "process power" as a plausible thesis rather than a proven financial driver.

Branding, network economies, counter-positioning — largely absent. Enterprise IT buyers outside South Korea do not pay a premium for the Samsung SDS brand name. Freight forwarding at this scale lacks meaningful network effects, as does managed cloud services. Moreover, Samsung SDS is the legacy incumbent being counter-positioned against by agile cloud hyperscalers, rather than the disrupter taking market share.

Four of Helmer's seven powers are present, with two strictly confined to a single captive parent. That combination creates a real moat, but one significantly narrower than corporate presentations suggest.

Porter's Five Forces

Buyer power splits cleanly along captive and non-captive lines. Within Samsung Group, affiliate buyer power is low in that subsidiaries cannot realistically migrate off core SDS platforms — yet high in that contract pricing is administratively capped by regulatory scrutiny of intra-group transactions and governed by a parent entity whose priorities do not always align with public shareholders. Outside the group, buyer power is unequivocally high: a commercial bank selecting a cloud provider evaluates multiple competitive offerings and retains significant pricing leverage.

Substitutes represent the most immediate threat in IT services. The primary substitute for a managed service provider is an enterprise customer building internal cloud capabilities or contracting directly with hyperscalers. As automated cloud migration tools advance, direct adoption becomes easier each year. In logistics, substitution risks are acute — freight forwarding remains a commoditized service where digital platforms like Flexport compete on software efficiency while incumbent global logistics providers bring immense scale and established carrier relationships.

Rivalry remains intense yet structurally static within South Korea. LG CNS, SK Inc. C&C, and Hyundai AutoEver each operate within their respective conglomerate ecosystems. A fresh public benchmark arrived in February 2025 when LG CNS listed on the KOSPI, pricing at KRW 61,900 per share for a market capitalization of approximately KRW 6 trillion and raising about KRW 1.2 trillion in South Korea's largest IPO since 2022.³²³³ Investment bankers used Samsung SDS as a primary valuation comparable.³² The stock debut was subdued, with shares falling 9.9% on day one.³² That performance underscores how public equity markets currently price the Korean IT-services model, indicating that the valuation discount facing Samsung SDS reflects broader sector dynamics rather than a firm-specific penalty.

New entrants encounter steep capital barriers in physical infrastructure alongside low barriers in software services. The principal competitive threat comes not from domestic IT startups, but from global hyperscalers deepening their direct presence in South Korea through local partner ecosystems.

Supplier power has emerged as a central operational constraint. In the enterprise AI infrastructure market, NVIDIA exercises extraordinary pricing power over hardware buyers. Samsung SDS has sought to mitigate this dependency by diversifying its silicon architecture — integrating domestic vendor FuriosaAI's Renegade NPU alongside GPUs on Samsung Cloud Platform for workload routing, while planning procurement across both NVIDIA's B300 and next-generation Vera Rubin architectures.³⁰ While this represents prudent supply-chain risk management, it does not alter the fundamental reality that input costs for AI compute remain dictated by a dominant vendor with far greater pricing power than its customers.

Where Samsung SDS actually competes, and where it does not

Synthesizing these dynamics reveals a clear operational picture: Samsung SDS is nearly unassailable within Samsung Group IT, competitive and growing in South Korean managed cloud and public-sector AI infrastructure, subscale globally in enterprise software, and a price-taker in international freight forwarding. For the long-term investment case to succeed, the non-captive cloud and AI segment must expand rapidly enough to shift the company's financial center of gravity. Its position in the captive IT foundation and low-margin logistics division will remain largely unchanged.

Executing that expansion depends on corporate leadership — bringing management strategy and executive execution into focus.

↑ Back to Top

X. Management Credibility, Governance & Risk Radar

Samsung SDS has been run by two very different technologists in the current era, and comparing them is instructive.

황성우 Hwang Sung-woo took the CEO role in 2021 from the leadership of the Samsung Advanced Institute of Technology, the group's blue-sky research arm, holding a doctorate from Princeton. He was a research scientist running a services business, and his tenure bore a scientist's signature: the Emro acquisition, the FabriX and Brity Copilot architectures, and a corporate posture focused on hyperautomation, which he outlined at length at the company's DTW 2024 conference in May 2024.³⁴

On November 28, 2024, Samsung SDS announced his successor: 이준희 Lee Jun-hee, an executive vice president from Samsung Electronics.³⁵ Lee's profile marked a clear operational shift. After earning degrees in electrical and electronic engineering at Seoul National University and MIT, he joined Samsung Electronics' DMC R&D Center in 2006, building his reputation commercializing the Galaxy smartphone line and Samsung's 5G network infrastructure.³⁶

That is not a blue-sky research background; it is an industrial shipping background focused on delivering hardware and network infrastructure at scale against strict deadlines. Appointing an executive with that operational record as the company prepared to commit billions of dollars to data centers, GPU procurement, and a national AI infrastructure contract signaled a clear shift in board priorities.

Judging management on behaviour, not statements

Three indicators offer a clear test of management's execution.

First, narrative consistency. The company's external cloud growth strategy has moved from vague targets to specific disclosures—progressing from broad non-captive expansion goals under prior leadership to named industry sectors, institutions, and explicit quarterly growth metrics under Lee Jun-hee. Struggling enterprises typically move in the opposite direction, toward vagueness, though Samsung SDS has still not disclosed the absolute revenue base for its external cloud business.

Second, transparency around earnings misses. The first quarter of 2026 tested management's communication when operating profit dropped sharply. Management clearly explained the decline by identifying and quantifying a KRW 112 billion one-time retirement benefit provision, while providing forward guidance on second-quarter growth drivers that proved directionally accurate.²³²⁴

Third—and where scrutiny remains most warranted—the capital allocation pivot. For a decade, leadership justified holding large cash reserves as balance-sheet prudence. In 2026, management declared the year pivotal, expanded its convertible bond ceiling more than twenty-fold, brought in KKR as a major convertible holder and six-year strategic adviser, and committed to an aggressive infrastructure expansion through 2031.²⁷²⁸³⁰ That represents a total strategic reversal.

While the pivot addresses long-standing investor criticism and targets a real AI compute opportunity, it also aligns with the current capital spending cycle. The true test of this capital allocation shift will be return on invested capital, which will not become visible until at least 2028.

The estate planning overhang, resolved

The governance uncertainty that shadowed Samsung SDS's first decade as a public company has largely closed.

Following the death of Chairman Lee Kun-hee, the controlling family faced an inheritance tax obligation of roughly KRW 12 trillion, paying it in five annual installments starting in April 2021 through affiliate share sales.³⁷ Samsung SDS shares were directly involved; in March 2022, Morgan Stanley and KB Securities placed approximately 3 million shares in a block trade attributed to Lee Boo-jin and Lee Seo-hyun.³⁸

The family settled the final tax installment on April 9, 2026, when 홍라희 Hong Ra-hee, Lee Kun-hee's widow, sold 15 million Samsung Electronics shares for approximately KRW 3.1 trillion, reducing her stake to 1.24%.³⁷

For public shareholders, this resolved the risk of recurring, mechanical share placements hanging over the stock. However, a broader structural overhang persists: family shareholdings mean Samsung SDS could still feature in future group reorganizations, a vulnerability highlighted by the contested 2016 split-off proposal.

Risk radar

Samsung Electronics capex sensitivity. Historically, about 40% of parent-entity revenue has come from Samsung Electronics alone.⁴ Because semiconductor capital expenditure is cyclical, a slowdown in fab construction directly delays SDS's infrastructure implementation revenue, with limited capacity to replace lost volume from other clients.

Logistics margin compression. This dynamic is already evident: 2025 logistics operating profit declined even as the segment generated more than half of consolidated revenue.² Normalizing freight rates are a permanent operating reality rather than a temporary disruption. The key question is whether Cello Square can build a high-margin software fee stream that isolates earnings from underlying shipping rates.

AI infrastructure over-capex. Committing to build more than 800 megawatts of data center capacity by 2031 represents a high-stakes bet on regional enterprise AI demand.³⁰ If adoption trails projections, Samsung SDS will be left holding costly, rapidly depreciating hardware with low utilization. Depreciation and amortization costs reached KRW 630 billion in 2025 before the main expansion began.⁶

Concentration in the national AI project. A consortium led by Samsung SDS—alongside Naver Cloud, Samsung C&T, 카카오 Kakao, Samsung Electronics, and KT—was selected to build South Korea's KRW 2.4 trillion to KRW 2.5 trillion National AI Computing Center in 해남 Haenam, designed to host 15,000 GPUs by 2028 and 50,000 by 2030, with construction breaking ground in early August 2026.³⁹⁴⁰⁴¹ While a major contract win, state-backed infrastructure projects carry policy exposure to government budget shifts and political changes.

Cybersecurity. This structural risk remains paramount. Samsung SDS's core value to Samsung Group hinges on operational trust and data security. A major security breach would not only impose financial costs, but also undermine the fundamental premise of its captive relationship.

Diworsification watch. Two 2026 transactions departed from the core enterprise strategy. In May 2026, Samsung SDS joined Samsung Securities and Samsung Card to acquire a combined 4% stake in 두나무 Dunamu—operator of the 업비트 Upbit crypto exchange—purchasing 1.39 million shares from Kakao affiliates for KRW 612.8 billion, with SDS taking a 1% stake.⁴²⁴³ In June 2026, the company backed U.S. robotics firm Walden Robotics via Samsung Venture Investment to explore robot orchestration for manufacturing execution systems.²⁴ While each investment has a logical justification, expanding into unrelated venture bets risks diluting focus from core enterprise cloud and AI infrastructure.

↑ Back to Top

XI. Investor Playbook, Key KPIs & Bull vs. Bear Case

Everything in this story compresses into a concise set of metrics for investors to monitor, along with a clear assessment of what must occur for the shares to re-rate—and what would invalidate the investment thesis.

The three KPIs that matter

1. External (non-Samsung Group) cloud revenue—both the growth rate and, critically, the absolute base. This metric underpins the entire re-rating case. A percentage growth rate without a base figure functions more as marketing than financial disclosure. Investors should track disclosed growth each quarter while pushing management for absolute revenue figures. If external cloud revenue compounds rapidly while expanding as a proportion of the company's broader cloud line—which exceeded KRW 2.7 trillion in 2025—Samsung SDS is successfully transitioning into an open-market cloud provider. If disclosure remains limited to percentage gains, that choice itself provides meaningful context.

2. The IT Services versus Logistics operating margin spread, and the profit mix. IT Services generates the overwhelming majority of operating profit on less than half of consolidated revenue. The investment thesis requires this earnings gap to widen through favorable mix shift—driven by higher-margin cloud services, SaaS offerings, and recurring AI deployments—rather than through logistics contraction alone. The critical test of capital efficiency will be whether IT Services operating margins expand as depreciation from the artificial intelligence build-out lands on the balance sheet.

3. Cello Square subscriber growth and its underlying unit economics. While subscribing client counts are regularly disclosed and growing, take rates and gross margins remain unquantified. The central analytical question is whether Cello Square operates as a genuine software platform earning high-margin fees on third-party freight volume, or primarily as a customer acquisition funnel for a thin-margin forwarding operation. Disclosure clarifying this economic structure represents a key catalyst for valuation.

The bull case

The bull case does not require Samsung SDS to evolve into a global software giant. The thesis is more targeted: Samsung SDS can establish itself as the premier enterprise AI infrastructure provider for a wealthy, technologically advanced, and structurally compute-constrained Korean market of 50 million people, prompting equity markets to value the firm beyond traditional systems integration multiples.

Four distinct pillars support this structural thesis.

First, domestic compute scarcity is acute, and Samsung SDS secured early access to advanced silicon while capturing initial public- and financial-sector contracts. Second, alliance frameworks across Anthropic, OpenAI, and Google Cloud position Samsung SDS as the core infrastructure layer regardless of which underlying model dominates enterprise adoption. Third, Samsung Group's internal ecosystem provides an immediate 70,000-seat deployment runway to validate products before commercial rollout. Finally, capital deployment is backed by an established global institutional partner with a financial incentive and formal advisory mandate to enforce balance-sheet discipline.

If external cloud revenue compounds through 2027 and 2028 while new AI capacity achieves high utilization, the resulting mix shift will fundamentally alter the consolidated earnings profile—enabling a business generating software-like incremental margins to decouple from freight-diluted trading multiples.

The bear case

The bear case does not depend on operational failure; it simply assumes recent historical patterns persist.

Samsung SDS has generated virtually flat consolidated revenue over a three-year period. Operating profit in 2025 reached KRW 957 billion compared to KRW 990 billion in 2019—representing six years of intensive strategic initiative without absolute earnings expansion.⁶ While the cloud segment has posted strong growth, it remains too modest a fraction of overall revenue to lift consolidated performance.

Three structural headwinds underpin the bear case. Captive IT revenue remains subject to regulatory caps and cyclical semiconductor capex. Logistics remains a commoditized forwarding business with compressed profit margins that distort consolidated metrics. Meanwhile, the managed service provider (MSP) business operates as an intermediary between hyperscalers and corporate clients—a position that historically faces margin compression as public platforms mature and enterprise buyers build direct integration capabilities.

The primary medium-term risk stems from capital allocation. Samsung SDS has committed to a multi-year, multi-hundred-megawatt infrastructure build at the peak of global AI infrastructure spending, funded partly through convertible instruments that dilute equity at KRW 180,000 per share. The company lacks a track record of managing capital expenditure at this scale. While depreciation expense is guaranteed, infrastructure utilization rates are not; if enterprise AI adoption lags projections, rising depreciation against underutilized assets will compress return on invested capital.

Finally, governance discounts tend to persist. Although inheritance tax liquidations are complete, the corporate structure that triggered past minority shareholder friction remains unchanged. Public investors in a chaebol affiliate hold an entitlement to underlying cash flows, but possess limited influence over structural corporate decisions.

Holding the two together

What makes Samsung SDS a compelling analytical case rather than merely a discounted stock is that the bull and bear arguments address two different operating realities. The bear case accurately describes the consolidated historical entity—flat, diluted, and structurally constrained. The bull case highlights a high-growth segment inside the business that possesses distinct technical assets, backed by committed capital and external strategic guidance.

The ultimate investment question is whether this high-margin cloud and AI business can expand rapidly enough to redefine the firm's overall valuation before committed capital costs become an unyielding drag on returns. That outcome will not be fully clear in 2026. However, the indicators that will determine it—external cloud disclosure levels, IT Services margins under heavy depreciation, and data center utilization across Dongtan, Gumi, and Haenam—will unfold quarter by quarter through verifiable operational metrics.

References

1. About Samsung SDS — Samsung SDS ↩↩↩↩↩

2. Samsung SDS Announces 2025 Financial Results — Samsung SDS, 2026-01-26 ↩↩↩↩↩↩↩↩↩↩↩↩

3. Samsung SDS Stock Quote & Financial Summary (018260.KS) — Reuters, 2026-08-09 ↩↩

4. Concerns rise over Samsung SDS' growing intra-group transactions — The Korea Times, 2023-06-04 ↩↩↩

5. Global Supply Chain: A competitive 4PL service based on the latest IT — Samsung SDS ↩

6. Samsung SDS Public Disclosures & Financial Statements — Samsung SDS ↩↩↩↩↩↩↩

7. Samsung SDS sets IPO price at 170,000-190,000 won a share — The Korea Herald, 2014-09-30 ↩

8. Samsung SDS in $1.1 Billion IPO — Cleary Gottlieb, 2014 ↩

9. Samsung SDS Surges as Investors Bet on Group Restructure — Bloomberg, 2014-11-14 ↩↩↩

10. Samsung SDS mulls split of logistics biz — The Korea Times, 2016-06 ↩

11. Samsung SDS faces fierce opposition over split-off plan — The Korea Herald, 2016-06-13 ↩

12. Three-year Dividend Policy (2025-2027) — Samsung SDS ↩

13. Samsung SDS Cloud Solutions — Samsung SDS ↩

14. Samsung SDS to Expand AI Infrastructure to 800MW by 2031, Accelerating Full-Stack Collaboration with Anthropic and Others — BigGo Finance, 2026 ↩↩

15. Top 5 Largest Data Center Projects in South Korea (2026) — Blackridge Research, 2026 ↩

16. Samsung SDS Signs 753 Billion Won in Dongtan Data Center Contracts — The Elec ↩

17. Samsung SDS Becomes the Largest Shareholder in emro, the Leading Procurement SCM Vendor in Korea — Samsung SDS, 2023-03-15 ↩↩↩

18. Samsung SDS Forms Business Partnership with o9 Solutions and emro to Gain Foothold in Global SaaS-based SCM Market — Samsung SDS ↩↩

19. Emro Appoints Samsung SDS Executive Vice President Sangwon Jo as New CEO — The Asia Business Daily, 2026-06-01 ↩

20. Samsung SDS Unveils Generative AI Services "FabriX" and "Brity Copilot" to Drive Hyperautomation in Corporate Business — Samsung SDS, 2024 ↩↩↩↩

21. Samsung SDS Adopts Anthropic's Claude and Forges Strategic Partnership to Expand Enterprise AI Business — Samsung SDS, 2026-07-25 ↩

22. Samsung SDS partners with Anthropic on enterprise AI — Korea JoongAng Daily, 2026-07-25 ↩↩

23. Samsung SDS Announces First-Quarter 2026 Financial Results — Samsung SDS, 2026-04-27 ↩↩↩↩↩

24. Samsung SDS Q2 Revenue Hits ₩3.72 Trillion on Cloud, Logistics Growth; External Cloud Business Surges 75% — BigGo Finance, 2026-07-30 ↩↩↩↩↩

25. Samsung SDS says second-quarter cloud revenue rises 17 percent, external business up 75 percent — Digital Today, 2026-07-30 ↩↩↩↩

26. Samsung SDS External Cloud Revenue Surges as B300 GPUaaS Goes Fully Commercial — Tech Times, 2026-07-30 ↩

27. Samsung SDS taps W6.4tr cash for AI, cloud push — The Korea Herald, 2026-03-19 ↩↩↩

28. Samsung SDS issues $820 mil. convertible bonds to KKR for future AI investment — The Korea Times, 2026-04-15 ↩↩↩

29. Samsung SDS shares jump 20% on KKR partnership and $820 million bond purchase — CNBC, 2026-04-15 ↩

30. Samsung SDS posts 2025 revenue of 13.93 trillion won, to push 'AI full stack' — Digital Today, 2026 ↩↩↩↩

31. Cello Square Conference 2025: Data-Based Logistics Innovation — Samsung SDS, 2025-03-28 ↩

32. LG CNS makes W6tr Kospi debut — The Korea Herald, 2025-02-05 ↩↩↩

33. Latham Advises on LG CNS's US$821 Million IPO on the Korea Exchange — Latham & Watkins, 2025-02 ↩

34. Samsung SDS President and CEO Sungwoo Hwang Unveils Future Vision of Generative AI and Hyperautomation Innovation at DTW 2024 — Samsung SDS, 2024-05-21 ↩

35. Samsung SDS Appoints New CEO — Samsung SDS, 2024-11-28 ↩

36. Samsung SDS names network specialist as new CEO — The Korea Herald, 2024-11-28 ↩

37. Samsung family completes inheritance tax payments with Hong Ra-hee share sale — The Korea Times, 2026-04-09 ↩↩

38. Samsung heiresses sell subsidiary shares to pay inheritance tax — The Korea Times, 2022-03-24 ↩

39. Samsung SDS Confirmed to Build Korea's National AI Computing Center — Seoul Economic Daily, 2026-05-11 ↩

40. Samsung SDS set to win bid to build South Korea's National AI Computing Center — Data Center Dynamics, 2026 ↩

41. Groundbreaking for the 2.4 Trillion Won National AI Center: "Haenam Becomes the Heart of Korea's AI" — The Asia Business Daily, 2026-08-03 ↩

42. Samsung units acquire 4% stake in Dunamu — The Korea Herald, 2026-05-28 ↩

43. Samsung Affiliates Acquire 4% Stake in Dunamu for 612.8 Billion Won — Seoul Economic Daily, 2026-05-28 ↩