Vessel AI Unveils Neocloud Growth Plan at Seoul Event

Vessel AI held a "Neocloud Strategy Presentation" in Seoul on October 15, 2026, where it announced mid- to long-term growth plans, according to techm.kr.

The company said it expects annual contract revenue of 250 billion won this year, 700 billion won next year, and more than 2 trillion won in 2028. It also plans to expand its GPU operating scale from 7,000 units to 10,000 by the end of this year and 50,000 next year, with a goal of 200,000 by 2030.

Vessel AI aims to expand data center capacity to 100MW by next year and 500MW by 2028. The company is in talks with domestic and foreign financial institutions to raise 4 trillion won to fund GPU purchases, and expects to need about 2 trillion won in financing for GPU expansion next year.

On the technology side, Vessel AI is running a project using more than 1,000 Nvidia B200 GPUs with SK Telecom. It will unveil a cluster based on Nvidia's next-generation DGX Vera Rubin NVL72 in the first half of next year. The company also plans to build an AI factory using domestic NPUs by 2028 and has completed a demonstration using Rebellions' Atom Max.

Vessel AI provides global inference infrastructure to Upstage's Solar Pro 4. The company started its business in 2020 developing machine learning operations software, and has been expanding its neocloud business since last year. CoreWeave, a competitor, invested large amounts of capital to secure data centers and GPUs.