This week, Broadcom (AVGO.US) highlighted its latest initiatives in AI agent governance, open-source security, and data sovereignty at the VMware Explore 2026 conference in Las Vegas. The semiconductor giant officially launched VMware Private AI Cloud, offering enterprises a production-ready path to securely build, run, and manage inference workloads, agent-based applications, and traditional enterprise workloads on a unified private cloud platform. This means businesses can now run AI production tasks in their on-premises environments.

Ram Velaga, President of Broadcom's Infrastructure Software Group, stated: "VMware Private AI Cloud represents a pivotal turning point, where enterprise private cloud and private AI infrastructure are no longer siloed but converge into a unified platform—enabling production-grade inference workloads and agentic AI to be deployed while meeting data sovereignty, compliance requirements, and cost predictability."

Additionally, Broadcom unveiled VMware AI Factory, serving as the software-defined foundation for VMware Private AI Cloud. Paul Turner, Chief Product Officer of Broadcom's VMware Cloud Foundation division, noted: "We provide customers with a software-defined underlying architecture that automates infrastructure deployment, unifies lifecycle management, and allows customers to choose their own hardware and validated models. The end result is faster deployment of the first model, predictable private cloud costs, and enhanced control over AI Tokenomics."

According to Broadcom, the AI Factory supports over 150 open-source and commercial models, is compatible with accelerators from NVIDIA (NVDA.US) and AMD (AMD.US), and has received certifications from multiple server manufacturers including Cisco (CSCO.US), Dell (DELL.US), Super Micro (SMCI.US), and Lenovo (00992). Suresh Andani, Corporate Vice President of AMD's Computing and Enterprise AI Business, remarked: "Combining AMD Instinct MI350 series GPUs with the open ROCm software ecosystem on VMware Cloud Foundation provides customers with a validated private AI path, featuring predictable cost structures without per-token billing models. This collaboration enables enterprises to run production-grade AI where their data resides while maintaining full-stack choice and control."

Furthermore, Broadcom released several new solutions for AI agent behavior control and governance, including AgentMinder, VMware vDefend, and the VMware Avi Load Balancer. Umesh Mahajan, Vice President and General Manager of Broadcom's Application Networking and Security Division, stated: "Enterprises can now safely operationalize agentic AI with a defense-in-depth security framework, implement strict guardrails, and ensure comprehensive protection of their AI environments."

Broadcom CIO Alan Davidson shared internal practices: "With the support of AgentMinder, our platform has achieved stable global-scale operations, maintaining zero downtime even during maintenance and upgrades. Through 'chain-of-custody supervision' between developers and multi-agent/skills, it has driven the full deployment of our company-wide AI agent pipeline. Its secure multi-region redundant architecture provides high-availability, low-latency connectivity services for over 20 million customer identities and 72,000 employee identities."

Broadcom estimates that 56% of enterprises are currently running or planning to run production-grade AI inference on private clouds. To align with this trend, the company also launched a new AI-ready data foundation for the VMware Tanzu platform. Tanzu, the official agent platform of VMware Private AI Cloud, offers hardened agent sandboxes, AI-ready data infrastructure, out-of-the-box developer tools, curated marketplaces, and auditable agent governance mechanisms.

On the financial front, Broadcom is scheduled to release its third-quarter earnings after the US market closes on September 2. Market consensus expects adjusted earnings per share of $3.24 on revenue of $29.43 billion, compared to adjusted EPS of $1.69 and revenue of $15.95 billion in the same period last year.