Today, at the AI in Finance Summit, Broadcom Inc. (NASDAQ:AVGO), a global technology leader that designs, develops, and supplies semiconductor and infrastructure software solutions, announced VMware Tanzu Platform agent foundations, introducing a secure-by-default agentic runtime designed to accelerate the delivery of autonomous AI applications. By extending the trusted code-to-production simplicity of Tanzu Platform to AI agents, Broadcom is enabling enterprise developers to move beyond siloed AI experiments and into scalable, governed production on VMware Cloud Foundation (VCF).

Bridging the Gap: From Experimental AI to Enterprise Reality

As AI agents begin to handle both software execution and autonomous decision-making, they require levels of governance and operational control that traditional platforms lack. Too often, organizations are forced to build AI projects in isolated environments that fail to integrate with core business data.

Tanzu Platform agent foundations solve this by providing a pre-engineered platform as a service (PaaS) environment for agents, built directly upon the trusted infrastructure layer of VCF. This allows platform engineers to manage AI services with the same tools they use for mission-critical business applications, without needing to become AI or data experts.

A Deny-by-Default Runtime for the Private Cloud

The new agentic runtime enforces a hard contract between developers and infrastructure, ensuring that agents remain within their authorized boundaries. Key innovations include:

  • Immutable Supply Chain: Instead of unverified Dockerfiles, the platform uses trusted Buildpacks to build agent containers automatically patched and verified, enabling automatic patching and eliminating the risk of embedded malware.
  • Structural Secrets Isolation: Prevents agents from reading each other's credentials at runtime – closing the door on lateral movement. Combined with VMware vDefend, that protection extends across infrastructure services and external SaaS connections.
  • Zero-Trust Networking and Sandboxing: limits runaway agentic loops through pre-defined resource limits. Connectivity to internal systems and models is never open by default. Access is explicitly granted only via secure service bindings, preventing wandering agents from accessing unauthorized data.



     

A Quick Start for Enterprise Developers

Tanzu Platform helps developers get started quickly, using a pre-built agent. Developers can give their agents governed access to models, Model Context Protocol (MCP) servers and marketplace services, all pre-curated by the IT organization. Agents can utilize integrated, enterprise-ready data engines, such as VMware Tanzu for Postgres with pgvector, caching, streaming, data flow services, and Spring AI memory services.

Day Two Operations: Scaling Agents on VMware Cloud Foundation

Tanzu Platform agent foundations programmatically leverage VMware Cloud Foundation IaaS APIs to abstract infrastructure complexity away from developers, enabling agents and their dependent services to always have the resources like compute, networking, and storage they need. Additionally, Tanzu Platform leverages VMware vSphere Kubernetes Service (VKS) to deliver scalable marketplace services.

  • Elastic Environment: Automatically scales up and scales down underlying IaaS resources to optimize cost and performance of short-lived and long-running agents.
  • High Availability and Lifecycle Automation: Provides four layers of high availability and self-heals infrastructure so that mission-critical autonomous applications remain resilient.
  • Model and Tools Serving and Brokering: Offers a centralized AI gateway to control tools and model availability, usage, costs, and safety filters across public models and private models on VCF.