Samsung SDS CEO NDR: AI data centers and KKR cooperation support a constructive view

Overview

This report summarizes Morgan Stanley's key takeaways following a non-deal roadshow with Samsung SDS CEO Junehee Lee. The discussion focused on AI data center expansion, cooperation with KKR, enterprise AI adoption, and the full-year earnings outlook. The report argues that GPU computing demand and data sovereignty trends are driving local data center construction, and Samsung SDS is well positioned to benefit through its AIDC capacity, group clients, and public-sector projects.

Core views

The core view is that stronger-than-expected demand for GPU racks and inference supports AIDC construction; the Dongtan West and Gumi projects improve visibility on data center revenue; enterprise AI remains at an early stage, with a multi-year installation cycle likely to unfold; the KKR partnership improves capital allocation and may drive ROE-accretive M&A; and an inflection in the memory market together with Samsung Electronics-related projects may boost a rebound in IT services revenue in 2H26.

Analysis framework

The report is based on management roadshow discussions, project capacity and contracting progress, data center economic assumptions, enterprise AI adoption trends, group client demand, and valuation frameworks, while breaking down upside and downside scenarios into demand, project contracting, M&A execution, and regulatory risk.

Methodology notes

• Valuation methodsSoTP + DCF

  sum-of-the-parts valuation and discounted cash flow

  The base case uses SoTP, with the core business valued using DCF, and includes an estimated equity NPV of Samsung SDS's stake in the consortium operating the National AI Computing Center project.

• Valuation assumptionsKey DCF parameters

  cost of capital, terminal growth, and net cash discount

  Key assumptions include WACC of 8.3%, beta of 0.9, risk-free rate of 3.5%, equity risk premium of 6.0%, a capital structure of 95% equity and 5% debt, a 50% discount to net cash, and a terminal growth rate of 2%.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Samsung SDS (018260.KS)

  The subject of the report, a Korean IT services and software company, with growth centered on AI data centers, enterprise AI services, cloud projects, and public-sector AI systems.

  Strengths

  AIDC capacity already shows clear contracting progress, Gumi DC has attractive target economics, the KKR partnership may improve capital allocation, and enterprise AI and public-sector demand are in the early stages of expansion.

  Weaknesses

  The data center business is highly capital intensive, expansion is constrained by power availability and capital limits, and revenue recovery still depends on key Samsung affiliates and the pace of government projects.

  Comparison

  Compared with traditional IT services, AIDC and enterprise AI services offer stronger growth and pricing flexibility, but also bring higher capital investment and project execution risk.

  Risks

  Slower demand for GPU computing services, weaker-than-expected progress in Gumi contract wins, ROE-dilutive acquisitions, regulatory issues for the National AI Computing Center, and weak IT budgets from Samsung affiliates.

• Samsung Electronics (005930.KS)

  A key affiliated customer and source of group demand; its business improvement, Taylor fab construction, and group AI process applications can support Samsung SDS's revenue recovery.

  Strengths

  An inflection in the memory market and broader adoption of AI applications at the group level can increase demand for IT services, cloud, and enterprise AI.

  Weaknesses

  If group IT budgets remain weak, Samsung SDS's revenue recovery and project demand may come under pressure.

  Comparison

  Group customer synergies improve demand visibility, but also increase dependence on capital expenditure and IT budget cycles within the Samsung ecosystem.

  Risks

  Budget cuts by Samsung affiliates, slower Taylor fab construction, or group AI deployment below expectations.

Key data

• Report date2026-06-14 11:05 PM GMTReport publication time disclosed in the main text.
• New capacity at Dongtan West20MWThe new capacity has been fully contracted, with revenue expected to ramp up in phases.
• Committed capacity at Gumi AIDC60MWThe company has secured 60MW of committed capacity and has the flexibility to add another 60MW.
• Target economics for Gumi DCHigh-teen operating margin; 10%+ unlevered IRRHigher token pricing may pose upside risk to earnings forecasts.
• Cash balance and KKR cooperationW8tn cash balanceKKR's convertible bond investment is viewed as a mechanism to drive more effective capital allocation and ROE-accretive investments.
• Equity valuation of National AI Computing CenterW1.8trn NPV; assuming SDS holds 33% of the SPCThis valuation is included in the SoTP base case.
• Latest historical target price recordW250,000 (2026-05-11)From the target price history in the report; the visible main text does not disclose the current price or expected upside.

Impact & implications

The report remains broadly constructive on Samsung SDS shares, arguing that tight supply-demand conditions for AI data centers, the enterprise AI installation cycle, and public-sector projects can improve medium-term growth visibility. If Gumi contracting, token pricing, and KKR-backed M&A are executed successfully, valuation upside may come from higher margins, IRR, and improved ROE; if Korean GPU computing demand slows, IT budgets from Samsung affiliates remain weak, or regulation of the National AI Computing Center is obstructed, the investment thesis would be weakened.

Risks

• IT budgets from key Samsung affiliates are weaker than expected.
• Growth in demand for Korean GPU computing services is slower than expected.
• Contract wins or revenue ramp-up at the Gumi data center fall short of expectations.
• The National AI Computing Center project encounters regulatory issues.
• KKR-backed acquisitions are delayed or cause value dilution.
• High capital expenditure and power supply bottlenecks limit data center expansion.
• The enterprise AI adoption cycle lengthens, and conversion of public-sector or enterprise projects is slower than expected.

What to watch

• Progress of the phased revenue ramp-up from the 20MW capacity at Dongtan West.
• Expansion and contracting status for an additional 60MW beyond the already committed 60MW at Gumi AIDC.
• Whether demand for GPU racks, localized data centers, and data sovereignty continues to exceed expectations.
• Actual delivery of Gumi DC operating margins, token pricing, and unlevered IRR.
• Investment reviews and the quality of M&A targets after KKR's participation as a board observer.
• The contribution of Samsung Electronics improvements, Taylor fab construction, and government projects to 2H26 revenue.
• Regulatory approvals for the National AI Computing Center project and progress in SDS's stake in the SPC.