KOSDAQIT & Software0158S0

엘리스

₩79,900▲ 29.92%2026-10-21 close

Market Cap

₩889.2B

Turnover

₩119.2B

Volume

1.6M

Shares out.

11.1M

PER

—

PBR

—

EPS

—

Dividend Yield

—

PER, PBR and dividend yield are calculated from the latest confirmed results (EPS, BPS, dividend per share) and the current share price · Prices as of the 2026-10-21 close

Report overview

IPO Stumble Tests Elice Inc.'s AI Infrastructure Pivot

Elice Inc. completed its KOSDAQ debut on October 20, 2026 after pivoting from edtech to AI full-stack infrastructure, but plunged 32.04% below its IPO price of ₩90,500 on its first trading day as the market demanded concrete proof of its AI infrastructure profitability narrative.

1
Shares of Elice Inc. closed at ₩61,500 on October 20, 2026 — down 32.04% from the IPO price of ₩90,500 — reducing the market capitalization to ₩684.4 billion versus the pre-listing target of ₩1.0054 trillion
2
H1 2026 consolidated revenue of ₩34.4 billion (up 202% year-over-year) and operating profit of ₩10.3 billion (operating margin 30.1%) mark a clear profitability recovery, with half-year revenue already equaling approximately 87% of full-year 2025 revenue
3
AI cloud revenue has structurally shifted from 6.6% of total sales in 2023 to 76.2% in H1 2026, confirming GPU-based cloud and data center services as the primary growth engine
4
The majority of the ₩201.1 billion IPO proceeds is earmarked for AI PMDC infrastructure expansion: approximately ₩126.3 billion for a new Honam-region facility and ₩28.7 billion for the Metropolitan-area site, with the rest targeting overseas markets
5
Near-term variables include concentrated government-sector revenue, a current ratio of 52% with short-term borrowings of ₩35.5 billion, and an additional wave of tradable shares expected approximately one month after the listing date

Business structure

Elice Inc. was founded in 2015 by alumni of KAIST's AI research lab and positions itself as an AI full-stack company delivering integrated AI infrastructure, cloud services, enterprise solutions, and education under a single entity. The business is organized into two main segments: AI Cloud and AI Transformation (AX).

The AI Cloud segment, which accounted for 76.2% of total revenue in H1 2026, is built around the proprietary portable modular data center 'AI PMDC' and the GPU virtualization platform 'ECI (Elice Cloud Infrastructure).' AI PMDC can be deployed in under three months — versus 24 to 36 months for conventional data centers — and reduces initial infrastructure build-out costs by up to 50%, constituting the core technological differentiator.

ECI is engineered to constrain GPU virtualization overhead to under 1%, and Elice Inc. holds Korea's first CSAP (Cloud Security Assurance Program) IaaS certification for a PMDC-based system, providing a meaningful competitive advantage in public-sector AI infrastructure procurement.

The AX segment operates the AI learning platform Elice LXP (over 2.8 million cumulative users), the enterprise generative AI 'Helpy' solution suite, and AI transformation consulting services, serving more than 2,000 paying customers including approximately 1,200 cloud clients.

Key domestic competitors include Naver Cloud and Samsung SDS — co-awardees with Elice Inc. on the government's large-scale GPU infrastructure project — while global hyperscalers such as Amazon Web Services, Microsoft Azure, and Google Cloud compete at the enterprise level.

Lead underwriter Mirae Asset Securities characterized Elice Inc. as a rare independent cloud service provider with actual GPU ownership in a domestic market otherwise dominated by large-conglomerate affiliates and asset-light managed service providers.

Recent trends

On a consolidated basis, Elice Inc. recorded fiscal-year 2025 revenue of ₩39.51 billion, operating profit of ₩5.18 billion, and net profit of ₩15.67 billion.

Operating profit declined for two consecutive years — from ₩13.2 billion in 2023 to ₩8.3 billion in 2024 and ₩5.2 billion in 2025 — driven primarily by a near-tripling of property, plant, and equipment additions (from ₩22.7 billion in 2024 to ₩63.8 billion in 2025), which significantly elevated depreciation charges.

The operating margin consequently compressed from 40.92% in 2023 to 13.12% in 2025. H1 2026 results marked a clear reversal: revenue of ₩34.4 billion and operating profit of ₩10.3 billion (operating margin 30.1%).

According to CNBC, the H1 2026 operating profit of approximately ₩10.3 billion represented a turnaround from an operating loss of approximately ₩3.5 billion in H1 2025, reflecting the operating leverage effect of rising infrastructure utilization rates.

H1 2026 EBITDA of ₩23.2 billion represented a 68% EBITDA margin; however, depreciation and amortization of ₩12.9 billion — equivalent to 37.4% of H1 2026 revenue — underscores the capital-intensity of the infrastructure model.

On the balance sheet, the current ratio stood at 52% as of end of H1 2026, with short-term borrowings of ₩35.5 billion and a debt-to-assets ratio of 28.72%, reflecting the strain of front-loaded capex.

Despite impressive IPO demand — a bookbuild competition ratio of 388.84:1 from 2,367 institutions and retail subscription of 408.26:1 with ₩10.2599 trillion in deposits — the market's first-day verdict was severe.

Outlook

Elice Inc. plans to deploy the bulk of its ₩201.1 billion IPO proceeds into AI data center infrastructure: approximately ₩28.7 billion for a four-building AI PMDC expansion at the Metropolitan-area site and approximately ₩126.3 billion for eight new AI PMDC buildings plus foundational network and storage infrastructure at a new Honam-region site.

CEO Jaewon Kim has indicated that government-led GPU projects are expected to wind down in H2 2026, with B2B commercial revenues projected to scale meaningfully from H2 2027.

On the international front, the company is pursuing a dual-track strategy — allowing overseas customers to remotely access domestic data center resources alongside physical AI PMDC deployments targeting the US, Japan, and Singapore.

The near-term challenge lies in demonstrating that the transition from concentrated government revenues to diversified private-sector and overseas revenues can be achieved on schedule.

A wave of additional shares is expected to become tradable approximately one month after the listing date, creating a potential near-term supply overhang.

Broader headwinds include deteriorating global investor sentiment toward AI infrastructure, particularly in the wake of profitability controversies surrounding US-listed CoreWeave, which may continue to weigh on the valuation framework applied to Korean AI infrastructure peers.

Bull factors

Explosive Cloud Revenue Growth and Emerging Operating Leverage

Annual cloud revenue of ₩3.0 billion in 2024 surged to ₩26.2 billion in H1 2026 alone, representing more than an eight-fold increase on a half-year basis relative to full-year 2024.

H1 2026 EBITDA margin of 68% and operating margin of 30.1% indicate that rising infrastructure utilization is now generating tangible operating leverage.

Track record on major government GPU infrastructure projects — including the ₩2.08 trillion program co-awarded with Naver Cloud and Samsung SDS — provides a credible reference base for future public and private contract wins.

The CSAP IaaS certification for PMDC-based infrastructure constitutes a meaningful entry barrier in Korea's government-designated cloud market.

AI PMDC Technology Edge and Independent CSP Positioning

AI PMDC enables data center deployment in under three months, cutting initial infrastructure costs by up to 50% versus conventional builds, providing a compelling time-to-market and total cost advantage.

By partnering with Arista Networks to implement large-scale Ethernet-based GPU clustering, Elice Inc. has moved beyond traditional InfiniBand-centric architectures to achieve both cost efficiency and scalability.

In a domestic market dominated by conglomerate-affiliated cloud providers and asset-light managed service providers, Elice Inc. is a rare independent, hardware-owning CSP with vertically integrated capabilities spanning data center construction, cloud operations, and enterprise AI solutions.

As the first independent Korean AI infrastructure company to list on KOSDAQ, the company has established a public-market disclosure and valuation benchmark for the sector.

Cross-Selling Potential from Education Customer Base and Global Footprint

The Elice LXP education platform, with over 2.8 million cumulative users and more than 2,000 paying customers, provides a natural cross-selling pathway from AI education clients to cloud infrastructure and AX solution services.

Conversion of existing education customers to higher-value cloud and AX contracts offers a capital-efficient route to revenue base expansion while lowering customer acquisition costs.

The Singapore digital textbook development project and planned expansion into the US, Japan, and Singapore establish a foundation for international revenue diversification.

Recognition as a '2026 Asia-Pacific High-Growth Company' by the Financial Times and Statista constitutes an independent third-party validation of the company's growth trajectory.

Bear factors

Heavy Government Revenue Concentration and Unproven B2B Transition

According to media reports, a single government entity accounted for approximately 66% of AI cloud revenue in H1 2026; as government-led GPU initiatives are expected to conclude in H2 2026, a potential earnings trough could emerge before private B2B revenues scale up — management's own projection places the B2B ramp at H2 2027 and beyond.

While the company has noted that most contracts under discussion carry three-to-five-year terms, the actual volume and timing of B2B conversions remain unverified.

If infrastructure utilization does not hold up through the government-to-commercial revenue transition, the high structural weight of depreciation would again compress operating margins.

In the education segment, the prior experience of AI digital textbook policy reversal — removal of mandatory adoption provisions from legislation — illustrates the business model's sensitivity to government policy shifts.

Balance Sheet Strain from Front-Loaded Infrastructure Investment

Capital expenditure surged to ₩63.8 billion in 2025, nearly 2.8 times the ₩22.7 billion spent in 2024, pushing H1 2026 depreciation and amortization to ₩12.9 billion — equivalent to 37.4% of H1 2026 revenue.

The current ratio stood at only 52% at end of H1 2026, well below the general 100% threshold, with short-term borrowings of ₩35.5 billion.

With approximately 81.5% of IPO proceeds committed to further facility investment, the company may require additional financing if utilization rates or new contract wins fall short of projections.

The structurally high depreciation load means operating margins are highly sensitive to even moderate shifts in data center utilization.

AI Sector Valuation Re-Rating and Post-IPO Supply Overhang

The IPO valuation was derived by applying an EV/EBITDA multiple of 27.13x to an annualized figure of ₩38.9 billion based on exceptionally strong H1 2026 EBITDA — the market's verdict on day one indicated skepticism about whether this growth rate is structurally sustained.

Global controversies over AI cloud companies' profitability, particularly CoreWeave's financial stability concerns, have directly dampened investor appetite for Korean AI infrastructure names.

Additional shares are expected to become tradable approximately one month after the listing, adding a persistent near-term supply overhang.

The compression from the pre-listing market cap of ₩1.0054 trillion to a first-day close of ₩684.4 billion signals an ongoing recalibration of the premium multiple applied to independent AI infrastructure operators.

Risk factors

Macro & Sector Risk

Global investor sentiment toward AI infrastructure turned more volatile in H2 2026, triggered by profitability controversies surrounding US-listed AI cloud companies.

Uncertainty regarding the timing and magnitude of returns on AI data center investment has not been resolved, leaving Korean AI infrastructure company valuations susceptible to shifts in the global technology investment cycle.

Changes in the interest rate environment can materially affect the financing costs and return profiles of capital-intensive business models requiring large-scale facility investments.

Elice Inc.'s overseas expansion plans targeting the US, Japan, and Singapore introduce additional risks related to local IT infrastructure policies, data sovereignty regulations, and foreign exchange rate movements.

Competitive Intensity Risk

Korea's AI cloud market features intense competition from large domestic conglomerates including Naver Cloud and Samsung SDS as well as global hyperscalers such as Amazon Web Services, Microsoft Azure, and Google Cloud.

Naver Cloud and Samsung SDS, which co-won the large-scale government GPU infrastructure project alongside Elice Inc., directly compete in the private B2B cloud market.

If global CSPs accelerate domestic data center investment in Korea or independently adopt modular data center approaches, Elice Inc.'s PMDC-based cost and deployment speed advantages could erode.

Transitions to next-generation GPU architectures and shifts in the AI semiconductor supply chain may also necessitate additional infrastructure reinvestment over the medium to long term.

Execution & Contract Risk

With approximately 81.5% of IPO proceeds committed to the new Honam-region data center, the company's financial performance is highly contingent on that facility achieving its targeted completion timeline, utilization levels, and contract wins.

A slower-than-expected B2B revenue transition or below-plan utilization would directly and negatively impact operating margins given the structural weight of depreciation in the cost base.

Historical precedent in the education segment — notably the removal of mandatory AI digital textbook adoption provisions — illustrates the vulnerability of policy-dependent business lines to regulatory shifts.

Should overseas revenue materialize later or at a smaller scale than management guidance implies, additional pressure for valuation re-rating could re-emerge.

Overall view

Elice Inc. has partially validated its edtech-to-AI-full-stack pivot through its H1 2026 financial results, with an AI cloud revenue share of 76.2%, operating margin of 30.1%, and EBITDA margin of 68% demonstrating that infrastructure utilization is translating into financial performance.

Nonetheless, concentrated government-sector revenues (with a reported single-entity dependency of approximately 66%), a current ratio of only 52%, an unproven B2B commercial transition pathway, and structurally elevated depreciation remain substantive concerns that the market priced in sharply on the first trading day.

The 32.04% decline against the IPO price reflects a confluence of global AI infrastructure valuation recalibration, near-term supply pressure from initial share unlocks, and deteriorating investor sentiment toward the AI cloud sector broadly.

At a current market capitalization of ₩684.4 billion versus the pre-listing IPO valuation of ₩1.0054 trillion, Elice Inc.'s ability to demonstrate sustained and diversified revenue generation — moving from government contracts toward private B2B and international customers — will be the defining metric for market participants.

The velocity of the private-sector B2B revenue transition following the conclusion of government GPU projects, and the utilization ramp of the upcoming Honam-region data center, are the key operational benchmarks to monitor through H2 2026 and into 2027.

Sources

Korea Exchange (KRX) — Prices · Market Cap · Volume
FSS electronic disclosure system (DART) — Financial statements · Dividend filings

Show 17 more articles and sources

Report written 2026-10-21 · Data as of 2026-10-20

This content is AI analysis of market data and web search results, provided for information only. It is not a solicitation or recommendation to invest. Investment decisions and their consequences are the investor's own responsibility. Data may be delayed or contain errors.