VESSL AI
Category: AI Infrastructure
Korean neocloud that rents GPUs to AI companies and runs model inference through its VESSL Cloud platform, built on its own 'Fluid Computing' orchestration. VESSL AI was founded in 2020. The company is led by Jaeman An. Based in Seoul, South Korea. Team size: 11-50. Total funding raised: $16.8M. Latest round: Series B. Key investors include A Ventures, KB Investment, Mirae Asset Venture Investment, Mirae Asset Securities, Ubiquoss Investment, Sirius Investment, SJ Investment Partners, Woosin Venture Investment, Shinhan Venture Investment, Springcamp.
- Founded
- 2020
- Headquarters
- Seoul, South Korea
- Team size
- 11-50
- Total funding
- $16.8M
Value proposition
Fast access to GPUs from a single L40S up to multi-node B200/B300/GB300 InfiniBand clusters. AI teams get preconfigured environments, per-second billing and inference with throughput SLAs, so they can focus on models instead of infrastructure.
Products and solutions
VESSL Cloud: Container Compute (Jupyter/SSH/IDE workspaces, batch jobs), VM Cluster (400Gbps NDR InfiniBand, NVLink, NFS, root access, managed Kubernetes/Slurm coming), Provisioned Throughput (OpenAI/Anthropic-compatible inference API with contractual SLAs), Storage. Also token-based inference service, legacy MLOps/LLMOps platform, and GPU supply contracts.
Unique value
Combines five years of MLOps software experience with owned GPU capacity. 'Fluid Computing' pools scattered GPU resources across data centers and clouds and allocates them to each workload.
Target customer
AI labs and model developers, enterprises training or serving models, university research labs, and overseas AI companies that need GPU capacity.
Industries served
Foundation model developers / LLM, enterprise AI, biotech/pharma, mobility/automotive, insurance/finance, academia, physical AI, image generation
Technology advantage
Proprietary Fluid Computing distributed-GPU orchestration; multi-cloud orchestration roots; latest NVIDIA Blackwell capacity (B200/B300/GB300); structured GPU financing with securities firms; planned network of 8 AI data center sites in Korea.
How they differentiate
Korea-based neocloud linking Korean data center capacity with global AI demand. Software-first orchestration heritage (MLOps). Offers both GPU rental and token-metered inference with SLAs, backed by financial-sector GPU financing.
Main competitors
CoreWeave, Lambda, Nebius (plus Crusoe and Korean cloud providers)
Key partnerships
Korea Investment & Securities (GPU financing MOU, Oct 2026), Neo AI Cloud (Pohang AI data center GPU platform), Upstage (global inference for Solar Pro 4), Google Cloud, Oracle, NVIDIA (GTC 2026 participant)
Notable customers
SK Biopharm, Upstage, Motif Technologies, LG AI Research, Naver, Hyundai Motor, TMAP Mobility, Hanwha Life, Scatter Lab, KAIST
Major milestones
2020-04 founded. 2021-11 ₩5B Pre-A. 2024-10 $12M Series A. 2026 pivot to GPU cloud/neocloud. 2026-09 supply deal for 128 NVIDIA B200 GPUs to SK Biopharm. 2026-09 Series C of about ₩200B reported in progress, led by KB Securities at about ₩500B pre-money, National Growth Fund considering participation. 2026-10 Series B results disclosed, 2026 Neocloud strategy event and ₩400B global GPU financing plan. 2026-10 GPU financing MOU with Korea Investment & Securities.
Growth metrics
Revenue was about ₩2.4B in 2024 and about ₩8B in 2025 (3.4x). H1 2026 revenue was ₩130B, with FY2026 forecast at ₩250B. Order backlog topped ₩100B as of Aug 2026. About 7,000 GPUs today; targets are 10K by end of 2026, 50K in 2027 and 200K by 2030. Plans for about 500MW of data center capacity by 2028.
Market positioning
Leading Korean neocloud, aiming to become a global top-tier neocloud. Moved from MLOps platform to GPU cloud provider in 2026.
Geographic focus
South Korea and the United States, plus global overseas AI customers
Patents and IP
Vessl AI Korea Inc. has published US patent applications including US20260140775A1, covering machine-learning operations in private networks, and US20260140704A1, covering LLM-based AI-agent generation. Both published May 21, 2026; these are applications, not confirmed grants.
About Jaeman An
Co-founded VESSL AI in April 2020. Named one of Forbes' Top AI Founders to Watch in 2025 and a Forbes Korea Rising AI Leader.
Latest news about VESSL AI
More AI Infrastructure companies
Official website: https://vessl.ai