NVIDIA is releasing deployment recipes optimized for its Dynamo platform to help bring four leading Korean open-source AI models into production environments, the company said Nov. 9, 2026. The recipes support A.X K2 from SK Telecom, K-EXAONE 2.0 from LG AI Research, Solar Open 2 from Upstage and Motif-3 from Motif Technologies. NVIDIA said the recipes address the challenges of deploying models that exceed the capacity of a single graphics processing unit. Such systems must distribute models across multiple GPUs, requiring coordinated request routing and other deployment controls. The release is part of NVIDIA’s effort to support production-scale use of open models developed by Korean technology companies and research organizations. Post navigation OpenAI Teams With Hacktron