Broadcom Launches AI-Ready Tanzu Data Foundation, Adding Deny-by-Default Controls for Enterprise Agents
Updated
Updated · InfoWorld · Sep 1
Broadcom Launches AI-Ready Tanzu Data Foundation, Adding Deny-by-Default Controls for Enterprise Agents
3 articles · Updated · InfoWorld · Sep 1
Summary
Broadcom said the new VMware Tanzu Platform foundation will reach general availability in the next few months, aiming to move enterprise AI agents from pilots into production.
Built directly into VMware Cloud Foundation, it keeps structured and unstructured data inside private clouds, adds semantic context, and lets developers deploy agents through prebuilt harnesses, memory, caching, and human-in-the-loop controls.
Deny-by-default sandboxes block network and API access until explicitly permitted, isolate credentials, and add lineage, prompt tracing, action attribution, and token-usage monitoring to curb data leaks, false outputs, and runaway costs.
IDC said the release avoids forcing customers onto a separate lakehouse by parsing data where it already resides, while supporting multiple frameworks and models inside the same governed sandbox.
Broadcom is pitching the platform at regulated sectors such as healthcare, insurance, and legal, while also targeting manufacturers and retailers that want AI behind their own firewall with data residency and auditability intact.
Could Broadcom's deny-by-default AI sandboxes be the secret weapon that finally protects autonomous enterprise agents from catastrophic data leaks?
What hidden performance costs lurk behind the heavy governance and MCP gateways required to keep enterprise AI agents strictly behind the firewall?
Will strict compliance standards and transaction-grade audit trails actually empower autonomous AI agents, or simply throttle them into glorified automation scripts?