AI infrastructure specialist Vessel AI (CEO Jae-man Ahn) announced on the 17th that it has advanced to the final stage of the 'Startup Grand Challenge 2025' hosted by NVIDIA Inception.
This challenge is a key event of the "N&UP" program, jointly run by NVIDIA, the Ministry of SMEs and Startups, and the Korea Institute of Startups and Entrepreneurship Development. It was designed to discover innovative startups in the AI and computing fields in Korea. Fifteen AI startups participated in the first live pitching round held on October 21st, and five companies, including Vessel AI, were selected as finalists through expert review and real-time audience voting.
Vessel AI will participate in the final pitching session at "NVIDIA AI Day Seoul," held at COEX in Seoul on the 17th and 18th. The company plans to present its "AI Orchestration Strategy," which integrates GPUs, models, and AI agents, as well as the data center and AI cloud operations that support it.
Vessel AI's core platform, "VESSL," is an orchestration infrastructure platform that integrates and manages GPUs, models, and AI agents across distributed cloud and data center environments. It supports major global clouds such as AWS, OCI, Azure, and GCP, and is designed to maintain high computational efficiency even in constrained environments.
Furthermore, by integrating NVIDIA's NIM (NVIDIA Inference Microservice) with the NeMo toolkit, efficiency from model development to deployment has been improved. By operating multiple models in parallel on the same server, cost savings of over 70% have been achieved on average. Furthermore, an automatic billing system based on GPU usage allows companies to flexibly scale their infrastructure.
- See more related articles
You must be logged in to post a comment.