NVIDIA pushes secure AI agents for telecom networks

NVIDIA is presenting a telecom autonomy stack at TM Forum’s DTW Ignite 2026 in Copenhagen, focused on moving operators beyond task-based automation toward AI agents that can monitor systems, coordinate changes and support autonomous network operations while keeping human policy controls in place.

Data remains a major constraint: 54% of operators cite data-related issues as their biggest barrier, with sensitive network and customer records often unavailable for model training. SoftBank Corp. is using NVIDIA NeMo Safe Synthesizer and NVIDIA NeMo Anonymizer to create privacy-preserving synthetic datasets for fine-tuning a large telecom model and building specialized network agents.

NVIDIA NemoClaw blueprints and NVIDIA OpenShell provide policy guardrails and sandboxed access for long-running agents. AdaptKey is piloting security-hardened agents for self-healing 5G operations, while Amdocs, NTT DATA, ServiceNow and TCS are applying the tools to customer care, migration planning, degradation detection, incident response and anomaly diagnosis.

Accelerated simulation is positioned as a safety layer before agents act on live systems. Forsk has integrated an AI-based radio propagation model into Naos, achieving ray-tracing-level accuracy up to 200x faster than CPU-only baselines on NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. KDDI and KDDI Research are also working with NVIDIA, Keysight and Samsung Research America on a high-fidelity RAN digital twin for 6G scenarios.