Vessel AI Advances to NVIDIA Startup Grand Challenge Finals

AI Infrastructure Innovation Company VESSL AI Achieves Feat of Advancing to Finals of NVIDIA 'Startup Grand Challenge 2025' VESSL AI (Vessel AI), a leader in the AI infrastructure sector, has proudly secured a spot in the finals of the prestigious 'Startup Grand Challenge 2025' hosted by NVIDIA Inception, solidifying its position as a company spearheading AI technology innovation in Korea. This challenge is a core event of the 'N&UP' program, jointly operated by NVIDIA, the Ministry of SMEs and Startups, and the Korea Institute of Startup & Entrepreneurship Development, aiming to discover and foster promising startups in South Korea's AI and computing sectors. In the first live pitching session held on October 21st, only five companies, including VESSL AI, earned the honor of advancing to the final round, having overcome fierce competition among 15 participating companies. VESSL AI will participate in the final pitching session at 'NVIDIA AI Day Seoul,' held from October 17th to 18th at COEX, Seoul, with a presenter on the 17th to showcase the company's vision and unique technological capabilities. Specifically, the company plans to present in depth its innovative 'AI Orchestration Strategy,' which organically integrates and manages GPUs, AI models, and AI agents, and how it leverages this strategy to operate optimal infrastructure in data center and AI cloud environments. This is anticipated to be a key solution capable of maximizing efficiency in complex AI development and operational environments. VESSL, VESSL AI's core platform, provides an advanced orchestration infrastructure that integrally manages GPU, AI model, and AI agent resources across distributed cloud and data center environments. It supports companies in effectively adopting and operating AI, and is designed to maintain excellent computational efficiency not only in major public clouds like AWS, OCI, Azure, and GCP but also in restrictive on-premise environments. In particular, the VESSL platform perfectly integrates with NVIDIA's NIM (NVIDIA Inference Microservice) and NeMo toolkit, dramatically improving efficiency across the entire process from AI model development to deployment. Furthermore, through an innovative structure that allows multiple AI models to operate simultaneously within the same server, companies can achieve infrastructure cost savings of over 70% on average. A flexible system that automatically calculates fees based on GPU usage supports companies in efficiently scaling and operating their AI infrastructure to match their business size. VESSL AI has already proven its technological prowess and stability by successfully supplying AI infrastructure solutions to leading domestic companies and public institutions such as Hyundai Motor Company, Tmap Mobility, Hanwha Life, and KAIST. Moreover, the VESSL platform is critically utilized in the process of optimizing learning and inference for major domestic large language models (LLMs) such as Naver's 'HyperCLOVA X', LG AI Research's 'EXAONE', and Upstage's 'Solar', significantly contributing to the development of the domestic AI ecosystem. VESSL AI is expected to leverage this advancement to the finals as a stepping stone to make a broader leap into the global AI infrastructure market.