AI News Feed
Market watch
Products & Applications

Cloudera launches Anywhere Cloud for secure AI agent access to data

Cloudera's new Anywhere Cloud platform gives AI agents governed access to sensitive data across clouds, on-premises, sovereign and edge environments.

The offering targets agentic AI workloads, in which autonomous agents perform work on behalf of humans. Unlike conventional chatbots, these agents require direct access to corporate data to complete tasks, but fragmented data environments and strict data sovereignty regulations have made that difficult. A recent Cloudera study showed that 73% of information technology leaders blamed infrastructure performance constraints for hindering their AI projects.

Chief Product Officer Leo Brunnick said in a statement to SiliconANGLE that enterprise AI has outgrown the public-cloud-only model. “Organizations shouldn’t have to choose between innovation and control,” he said. “Cloudera Anywhere Cloud brings the speed and flexibility of the cloud directly to enterprise data, allowing customers to accelerate AI initiatives while maintaining complete ownership of it, with optimized cost of infrastructure and tokens.”

Cloudera Anywhere Cloud uses a unified application programming interface and the open-standards Apache Iceberg table format to connect analytics engines to data in place, without custom integrations. Companies can deploy tools such as Apache Spark, Kafka or Trino alongside third-party tools. Requests can be made through a plain-language agent, which converts them into automated workflows. The platform also employs zero-trust governance to ensure compliance and data lineage.

Brunnick explained that the platform activates data where it lives, enabling low-latency querying for agents. For on-premises deployments, it can run a fine-tuned LLM locally; for sovereign clouds, it operates within the sovereign ecosystem; and for edge environments, it runs across near and far edge devices. In multi-cloud scenarios, Anywhere Cloud can federate data when latency is less critical or replicate and stream data when low latency is required. Some early adopters have already built multi-agent workloads that co-locate agents in multiple data environments before reasoning across clouds and data centers.