Cloudera has announced on-premises AI inference capabilities powered by NVIDIA technology, signaling an expansion strategy that integrates AI model deployment, analysis, and visualization in the data center.

Cloudera announced on February 10, 2026 in Seoul that it has unveiled the on-premises expansion of its Cloudera AI Inference and Trino-based data warehouse.

As companies move beyond AI experimentation and into operational use, the company explains, there's a growing need for AI to be able to securely access data regardless of its location. Citing its own report, Cloudera added that approximately 42% of all companies store their data in data warehouses.

Cloudera AI Inference, available on-premises, is designed to enable enterprises to deploy and scale models for LLM, anomaly detection, computer vision, and speech AI within their data centers. The company said it supports a variety of models, including the latest NVIDIA Nemotron open model.

Cloudera explained that this feature is accelerated by the NVIDIA AI Stack, Blackwell GPUs, Dynamo-Triton inference servers, and NIM microservices, focusing on avoiding cloud cost volatility and controlling latency, regulatory compliance, and data privacy.

Cloudera Data Warehouse, based on Trino, announced that it will provide centralized security, governance, and visibility with expanded data center support, and improve access to insights by integrating AI-based analytics and visualization capabilities.

Data Visualization updates include AI Annotations, which automatically generates charts and visual data summaries, AI query logging and tracking, and improved management capabilities, including single sign-on (SSO)-based configuration. Cloudera explained that this expansion is designed for scenarios where AI and analytics can be run in the data center without data movement.