Korea’s AI Ecosystem Advances and Secures Agentic AI With NVIDIA

NVIDIA says several South Korean organizations are building on its software for AI agents and open models, according to a company blog post by Nicola Sessions dated November 8, 2026. The announcements cover two fronts: keeping autonomous agents secure inside businesses and giving developers models tuned to Korean contexts.

On security, NVIDIA reports that Hyundai Motor Group is exploring NemoClaw blueprints and the OpenShell runtime for agents aimed at employee productivity and enterprise workflows. OpenShell, part of what NVIDIA calls its Open Agent Safety Platform, runs agents in sandboxes and applies policy-based controls intended to protect sensitive data, manage access to company tools and govern what agents do on their own. Korean AI security firm AIM Intelligence is integrating its Starfort agent control platform with OpenShell, pairing the runtime's access controls with Starfort's detection of malicious instructions and data leakage. NVIDIA adds that AIM Intelligence is considering a preliminary proof of concept in the public energy sector, with possible later uses in enterprise, industrial and robotics settings.

On models, LG AI Research is opening API access to K-EXAONE-2.5, which was developed using NVIDIA's Nemotron datasets, including Nemotron-Personas-Korea. That open dataset contains 7 million fully synthetic Korean personas based on official national statistics, and model builders use it to ground training data and evaluation in local contexts.

The post is a vendor announcement, and much of the activity it describes remains exploratory, particularly Hyundai's work and AIM Intelligence's energy pilot. Still, it shows NVIDIA positioning its open tools as a foundation for both agent security and Korean-language AI development.