Trend Micro

Solution

Trend Micro uses its proprietary Trend Cybertron cybersecurity LLM together with the NVIDIA universal LLM NIM microservice to operationalize secure AI inference on NVIDIA Enterprise AI factories. The implementation is built around security-specific inputs including risk intelligence, customer telemetry, adversarial tactics, MITRE ATT&CK techniques, SIEM logs, natural-language risk reports, red-team narratives, security knowledge bases, and threat intelligence. The universal LLM NIM microservice serves Cybertron in cloud, hybrid, and on-premises deployments and abstracts inference backends so Trend can run NVIDIA TensorRT-LLM for high-performance inference or vLLM for throughput and batching. Trend pairs this serving layer with the AI-Q Blueprint to build RAG-style security agents that reason over knowledge bases and telemetry in real time, while NVIDIA NeMo Data Flywheel is used to monitor inference behavior and retrain models to manage drift and sustain accuracy.

Data flow

Trend feeds Trend Cybertron with security data such as risk intelligence, telemetry, MITRE ATT&CK techniques, SIEM logs, risk reports, red-team narratives, knowledge bases, and threat intelligence. Cybertron is served through the NVIDIA universal LLM NIM microservice on NVIDIA Enterprise AI factories, using TensorRT-LLM or vLLM as the inference backend. AI-Q Blueprint then enables agents to reason over those live security sources, while NVIDIA NeMo Data Flywheel monitors production behavior and retrains models as needed.