| Mobile Web

On-premise LLMs must approach cloud performance to succeed in enterprises

South Korean software company KMS Lab is working on POSCO’s next-generation enterprise portal using its AI-based knowledge platform, K-Portal BrainWeave. CEO Hee-jeong Bae said the company supplied a full-stack AI solution, from model serving and training data processing to business applications. The on-premise project runs LLMs on in-house GPUs without sending data to external clouds. Bae said enterprises will not use on-premise models unless performance approaches cloud LLMs.