StocksMedium•20 October 2026•
3 min read

Elice shares fall 32% on first morning of Kosdaq trading despite strong IPO demand

Key Facts

1Elice shares fell 32% from their offer price to 61400 won during morning trading on their Kosdaq debut.
2Elice set its offer price at the top of the indicative range at 90,500 won per share.
3A total of 2,367 institutions participated in bookbuilding, with institutional demand reaching 388.8 times the shares on offer.
4Elice reported consolidated revenue of 344 hundred million won and operating profit of 103 hundred million won in the first half of 2026.

Shares in Elice, a provider of cloud services and artificial intelligence infrastructure, fell 32% from their offer price in Tuesday morning trading on their Kosdaq debut, according to a report published by Investing.com. The shares traded as low as 61400 won, marking a weak start for the company as it seeks to expand its data centers. Elice had set its offer price at 90,500 won per share, according to its final pricing statement. The price was at the top of the indicative range after bookbuilding attracted strong institutional demand. The Investing.com report linked weak market appetite for the AI infrastructure sector to concerns over borrowing costs and long-term returns on massive spending.

According to Elice's announcement, 2,367 institutions participated in bookbuilding, with institutional demand reaching 388.8 times the shares on offer. The company said IPO proceeds totaled about 2,011 hundred million won. It estimated its market capitalization at listing, based on the offer price, at about a trillion and 54 hundred million won. Mirae Asset Securities led the offering, with Samsung Securities participating, according to the listing application statement. The stock's morning decline contrasted with strong demand during the IPO process, which ended with pricing at the top of the announced range.

On earnings, Elice reported consolidated revenue of 344 hundred million won in the first half of 2026. It posted consolidated operating profit of 103 hundred million won for the same period, according to the company's statement on its Kosdaq offering. The company said first-half revenue alone equaled 87% of revenue for the entire previous year. First-half earnings before interest, taxes, depreciation and amortization totaled 232 hundred million won. The company added that its EBITDA margin reached 68% as its AI infrastructure business expanded.

Elice's operations include cloud services and AI transformation solutions, according to its listing application statement. Its business thus combines computing infrastructure with solutions for deploying AI within organizations. In data centers, the company uses a modular model involving the prefabrication of equipment in a factory, according to a separate statement. This equipment includes the servers, power systems, cooling systems and networks used to build a data center. Elice also secured strategic investment from Dongkuk Holdings and GS Ventures, which it said supports cooperation in AI infrastructure.

Chief Executive Jae-won Kim said in the strategic investment statement that "AI infrastructure is a core element directly linked to national competitiveness." The investment from Dongkuk Holdings and GS Ventures was intended to support cooperation in this field, according to the company. Meanwhile, the Investing.com report pointed to market concerns over the cost of financing AI infrastructure. It also linked those concerns to long-term returns on massive spending in the sector. The report cited Firmus's cancellation of its planned Sydney listing during the same month as another example of setbacks to IPO plans tied to the industry.

Elice plans to use the IPO proceeds to expand its edge AI data centers, according to the Investing.com report. Its plans include building more data centers in South Korea and abroad. Its overseas expansion is linked to its entities in the United States, Japan and Singapore, according to the report. Alongside those plans, the company is participating in a government project to provide computing resources for scientific research in AI. Elice said the provision of high-performance graphics processing units, essential for large-scale AI research, runs from July this year through next March.