Broadcom adds AI security tools to VMware Tanzu Platform

Broadcom has added new AI data management and security features to VMware Tanzu Platform, targeting companies running AI agents inside private clouds.

The release is intended to help enterprises move from pilot projects to operational AI systems while keeping tighter control over software agents and the data they use. Broadcom is positioning Tanzu Platform as the agent platform for VMware Private AI Cloud.

At the centre of the update is an effort to address what Broadcom describes as the trust problem around agentic AI. Unlike conventional software, AI agents can act with a degree of autonomy and query data sources directly, raising concerns over data leakage, inaccurate outputs, and rising cloud costs when workloads spill outside private environments.

The latest version of Tanzu Platform tackles that issue in two ways: by securing the agent itself and by governing the data it accesses. In practice, agents run in isolated sandboxes with tightly limited permissions, while data controls are designed to restrict access, provide context, and preserve lineage so users can see what information was used and where it came from.

New controls

The additions include hardened agent sandboxes built around a deny-by-default model, intended to isolate credentials and reduce the risk of prompt injection attacks or unauthorised network access. Broadcom has also introduced what it calls AI-ready data foundations, which process structured and unstructured enterprise data on site before AI agents use it.

That on-site data preparation is meant to improve the relevance of information supplied to models while reducing token use. The system is also designed to improve output accuracy and make it easier for companies to keep sensitive information inside their own infrastructure rather than sending it to external services.

The release also includes an out-of-the-box developer harness with pre-approved skills, workflow buildpacks, human-in-the-loop controls, and integrated memory services. A curated marketplace has been added as a central catalogue for vetted AI models, tools, and data products that developers and agents can connect to within the platform.

Another feature, auditable agent governance, adds an AI gateway designed to monitor, rate-limit, and log each action an agent takes. That could appeal to organisations facing strict audit and compliance demands as they move beyond limited AI trials.

Enterprise demand

The announcement reflects a broader shift in enterprise AI adoption. Many businesses have spent the past two years testing generative AI in contained pilots, but moving those systems into production has exposed harder questions around oversight, infrastructure control, and the quality of the data feeding automated decisions.

Rachel Stephens, Research Director, RedMonk, said those concerns often extend beyond obvious risks. “Enterprises aren't hesitant about agents because of what they know can go wrong; it's the unknown unknowns, the unintended consequences nobody can enumerate in advance. Much of that risk lives where agents meet enterprise data. Platforms that make data access governed, curated, and auditable by default turn that fear into something manageable.”

Broadcom argues that private cloud environments offer one path for companies that want to use AI systems without ceding control over data or infrastructure. It is also seeking to tie that argument to Tanzu, which has traditionally been associated with modern application deployment and Kubernetes-based operations.

Purnima Padmanabhan, Vice President and General Manager, Tanzu Division, Broadcom, said the barrier is not a lack of appetite for AI. “Enterprises do not have an AI ambition problem. They have an agent trust problem. You cannot let software make decisions on data you do not govern, running on infrastructure you do not control. Tanzu Platform for Agents pairs governed agents with governed data inside the customer's own boundary, so businesses can move fast with agents without giving up control.”

User example

Broadcom also pointed to existing use of Tanzu in enterprise application operations as a foundation for broader AI adoption. One customer, Business Development Bank of Canada, said standardising deployments across teams reduced manual processes and eased the burden on internal platform staff.

Marie-Claude Potvin, Assistant Vice President, Infrastructure Services, Business Development Bank of Canada, said: “As part of an integration phase that began in 2019, BDC used VMware Tanzu Platform to support the deployment and operation of selected internal business applications. During this phase, the platform helped standardize how applications were deployed and managed across teams, reducing manual processes and differences in how applications were run. As our use of data and AI evolved, this also helped our platform teams support these applications and related services without requiring additional specialized expertise. This helped reduce day-to-day operational overhead for our technology teams.”