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

Agents that can call tools and move data are only as trustworthy as the box they run in. Hyundai Motor Group is exploring NVIDIA's NemoClaw blueprints and the OpenShell runtime, and Korean security firm AIM Intelligence is integrating its Starfort agent-control platform with OpenShell.

On the model side, LG AI Research is opening API access to K-EXAONE-2.5, built using NVIDIA's Nemotron datasets including a Korean synthetic persona set. Together, it's a small but useful sign that secure deployment is becoming part of the agent pitch rather than an afterthought.

We're encouraged by the order of operations here. OpenShell provides sandboxed execution and policy-based controls over data, tool access and agent actions. AIM Intelligence wants to layer detection of malicious instructions and data leakage on top. Runtime limits plus detection is the right shape: one stops an agent from doing what it shouldn't, the other notices when someone is trying to make it.

The caveats are real. Hyundai is exploring, not deploying. AIM Intelligence is considering a preliminary proof of concept for the public energy sector. That's early, and energy is exactly the kind of setting where a compromised agent could do physical harm. Sandboxes are only as good as the policies written into them, and prompt-injection detection has a long history of being bypassed. I'd want to see red-team results before anyone calls this secured.

The LG piece deserves its own look. Nemotron-Personas-Korea is an open set of 7 million fully synthetic Korean personas grounded in official statistics. That's a sensible way to make a model fluent in local context without scraping private data. The risk with synthetic personas is subtler: they can flatten a population into its averages, and anything the statistics miss, the model may miss too. Evaluations built on the same synthetic base can make that blind spot hard to see.

OpenShell is part of what NVIDIA calls its Open Agent Safety Platform. Nemotron is NVIDIA's family of open models, datasets and techniques for building specialized agents. These items were announced alongside other NVIDIA partnerships in Korea covering startups and physical AI.