TLDR
- Broadcom unveiled VMware AI Factory at VMware Explore 2026, promising to cut bare metal deployment time from weeks to hours
- The platform supports over 150 open-source and commercial AI models and works with GPUs from Nvidia and AMD
- New AI agent governance tools were announced, including AgentMinder, VMware vDefend, and VMware Avi Load Balancer
- Broadcom says 56% of enterprises are already running or planning production AI on private cloud
- Q3 earnings are due Wednesday Sept. 2, with analysts expecting EPS of $3.24 on revenue of $29.43B
Broadcom dropped a wave of announcements at VMware Explore 2026 in Las Vegas on Monday, centering on private AI infrastructure and agent security.
The headline product is VMware AI Factory, the software-defined core of VMware Private AI Cloud. It is designed to let enterprises run AI workloads on their own infrastructure rather than relying on public cloud providers.
AVGO stock was down 0.08% on Monday as the news hit.
The pitch is speed and cost control. Broadcom says VMware AI Factory can take a server from bare metal to running its first AI model in hours rather than weeks. That is a meaningful reduction for enterprise IT teams managing large deployments.
The platform pools GPU resources across teams so organizations share hardware instead of dedicating separate infrastructure to each workload. Broadcom frames this as a way to keep costs in check and improve what it calls “AI tokenomics.”
Broad Hardware and Model Support
VMware AI Factory works with accelerators from Nvidia and AMD. On the AMD side, Broadcom confirmed support for AMD Instinct MI350 Series GPUs paired with the open ROCm software ecosystem. AMD’s corporate vice president Suresh Andani said the setup gives customers “predictable economics and no per-token pricing.”
Certified server partners include Cisco, Dell Technologies, Lenovo, and Supermicro. The factory supports more than 150 AI models including Nemotron 3, Gemma 4, Qwen 3.7-Max, and GLM 5.2.
Broadcom also announced a new partnership with MetalSoft to handle bare metal automation, cutting physical server provisioning from weeks to minutes directly through the VCF management console.
New private AI features include multi-tenant model sharing through isolated namespaces, an AI Gateway for unified model governance across on-premises and cloud, and secure AI sandboxes that isolate agent-generated code execution.
AI Agent Governance Takes Center Stage
A notable chunk of the event focused on governing AI agents, which is becoming a growing concern for enterprises running agentic pipelines.
Broadcom introduced AgentMinder, which provides chain-of-custody controls between developers and multiple agents. Broadcom CIO Alan Davidson said it supports over 20 million customer identities and 72,000 workforce identities with zero downtime.
VMware vDefend and VMware Avi Load Balancer were also announced as part of the agent security stack. Broadcom said enterprises can now enforce guardrails and protect AI environments with what it calls “defense-in-depth security.”
Broadcom cited internal research showing 56% of enterprises are already running or planning to run production AI on private cloud, which it says justifies the platform’s direction.
Q3 earnings are due after market close on Wednesday, Sept. 2. The consensus estimate calls for adjusted EPS of $3.24 on revenue of $29.43B. That compares to adjusted EPS of $1.69 on revenue of $15.95B in the same quarter last year.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.







