Broadcom (AVGO) Stock Eyes Q3 Earnings as VMware AI Factory Debuts

Key Highlights

• VMware AI Factory made its debut at VMware Explore 2026, designed to reduce bare metal server deployment from weeks down to just hours
• The new platform provides compatibility with more than 150 AI models, both open-source and commercial, and integrates with GPU hardware from Nvidia and AMD
• Broadcom rolled out AI agent security solutions including AgentMinder, VMware vDefend, and VMware Avi Load Balancer
• Internal research from Broadcom indicates 56% of enterprise organizations are either running or preparing to deploy production AI workloads on private cloud infrastructure
• Third-quarter financial results are scheduled for release Wednesday, Sept. 2, with Wall Street anticipating $3.24 earnings per share on $29.43B in revenue

At VMware Explore 2026 in Las Vegas on Monday, Broadcom unveiled a series of major announcements focused on private AI infrastructure and security for AI agents.

Broadcom Inc., AVGO

The centerpiece announcement is VMware AI Factory, which serves as the software-defined foundation of VMware Private AI Cloud. The solution enables businesses to execute AI workloads on proprietary infrastructure instead of depending on public cloud services.

Shares of AVGO declined 0.08% on Monday following the announcement.

The value proposition focuses on deployment velocity and financial efficiency. According to Broadcom, VMware AI Factory can accelerate server provisioning from bare metal configuration to running initial AI models in a matter of hours instead of the traditional weeks-long timeline. This represents a substantial advantage for enterprise IT departments handling extensive deployments.

The system consolidates GPU computing power across different teams, enabling organizations to utilize shared hardware infrastructure rather than allocating dedicated resources for individual workloads. Broadcom positions this approach as a strategy to manage expenses and enhance what the company describes as “AI tokenomics.”

Extensive Hardware and AI Model Compatibility

VMware AI Factory integrates with GPU accelerators from both Nvidia and AMD. For AMD integration, Broadcom verified compatibility with AMD Instinct MI350 Series GPUs combined with the open-source ROCm software platform. AMD’s corporate vice president Suresh Andani highlighted that this configuration delivers “predictable economics and no per-token pricing” for customers.

The platform has received certification from major server manufacturers including Cisco, Dell Technologies, Lenovo, and Supermicro. VMware AI Factory provides access to over 150 AI models such as Nemotron 3, Gemma 4, Qwen 3.7-Max, and GLM 5.2.

Broadcom also revealed a strategic collaboration with MetalSoft to streamline bare metal automation, reducing physical server setup time from weeks to minutes through direct integration with the VCF management console.

Additional private AI capabilities include multi-tenant model distribution via isolated namespaces, an AI Gateway offering unified model oversight across on-premises and cloud environments, and protected AI sandboxes that contain agent-generated code execution.

Enterprise Focus Shifts to AI Agent Management

A significant portion of the conference addressed AI agent governance, reflecting mounting enterprise concerns around managing agentic workflows.

Broadcom launched AgentMinder, delivering chain-of-custody oversight between developers and multiple AI agents. Broadcom CIO Alan Davidson noted the system manages more than 20 million customer identities and 72,000 workforce identities while maintaining zero downtime.

VMware vDefend and VMware Avi Load Balancer were unveiled as components of the agent security infrastructure. According to Broadcom, enterprises can now implement protective guardrails and secure AI environments through what the company characterizes as “defense-in-depth security.”

Broadcom referenced proprietary research revealing that 56% of enterprises have either deployed or are planning to deploy production AI on private cloud infrastructure, validating the platform’s strategic direction.

Third-quarter earnings results are scheduled for release after market close on Wednesday, Sept. 2. Wall Street consensus projects adjusted earnings per share of $3.24 on revenue of $29.43B. This compares to adjusted EPS of $1.69 on revenue of $15.95B during the corresponding quarter in the prior year.