VMware Cloud Foundation 9.1 introduces a unified private cloud platform for production AI, combining Kubernetes, security, and multi-vendor GPU support to help enterprises deploy AI workloads more efficiently and cost-effectively.
 |
| VMware Cloud Foundation 9.1 delivers a secure, cost-optimized, AI-native private cloud platform with Kubernetes integration, multi-GPU support, and automated operations for enterprise-scale AI deployments. Image: CH |
Tech Desk — May 19, 2026:
Broadcom has introduced VMware Cloud Foundation (VCF) 9.1 as a unified private cloud platform designed specifically for the growing demand of production AI workloads. Built on VMware Cloud Foundation, the new release focuses on reducing infrastructure costs, improving security, and simplifying the deployment of AI inference and agentic applications across enterprise environments.
At its core, VCF 9.1 is designed to help organizations move AI from experimentation into production by running AI workloads, containers, and virtual machines on a single, integrated infrastructure layer. Instead of managing separate systems for Kubernetes, VMs, storage, and GPUs, enterprises can now operate everything through one unified platform optimized for AI-scale performance.
To understand how to use VCF 9.1 effectively, it helps to break it down into its operational model: deployment, workload management, optimization, and security.
The first step in using VMware Cloud Foundation 9.1 is building a private cloud environment that can support both traditional enterprise applications and modern AI workloads.
VCF 9.1 supports an open hardware ecosystem, meaning enterprises can deploy on AMD, Intel, and NVIDIA-based infrastructure without locking into a single vendor. This flexibility allows organizations to design clusters depending on workload requirements, whether CPU-heavy agentic workflows or GPU-accelerated inference systems.
The platform is also Kubernetes-native, which means AI applications can be deployed using containerized workflows alongside traditional virtual machines. This eliminates the need for separate infrastructure stacks and allows IT teams to standardize deployment processes across AI and non-AI workloads.
Once the platform is deployed, VCF 9.1 enables enterprises to run production AI workloads, including inference systems and agent-based applications.
A key design principle in VCF 9.1 is mixed compute support. Many AI systems today rely heavily on GPUs for inference, but agentic AI workflows often require significant CPU resources for orchestration, tool usage, and decision execution. VCF 9.1 is built to manage both workloads simultaneously on the same infrastructure.
This unified approach helps reduce operational fragmentation and allows enterprises to scale AI applications without building separate GPU-only clusters.
According to Broadcom, the platform can deliver up to 46% reduction in Kubernetes operational costs and significantly improve deployment speed and scalability for AI workloads, making it suitable for production environments where uptime and performance are critical.
One of the central advantages of VCF 9.1 is infrastructure optimization, particularly important in an era of rising AI compute costs and hardware shortages.
The platform uses intelligent memory tiering and storage compression techniques that can reduce server costs by up to 40% and storage total cost of ownership by up to 39%. These optimizations allow enterprises to run more AI workloads on existing infrastructure instead of constantly expanding hardware capacity.
In addition, automated fleet management improves operational efficiency by enabling faster cluster upgrades and scaling across distributed environments. This is particularly important for organizations running large-scale AI systems across multiple data centers or hybrid cloud environments.
For IT teams, this means less manual intervention and more predictable scaling as AI demand grows.
Security is a core component of VMware Cloud Foundation 9.1, especially as AI workloads increasingly involve sensitive data, proprietary models, and regulatory constraints.
VCF 9.1 is built around a zero-trust architecture that includes continuous compliance monitoring, ransomware recovery tools, and secure workload isolation. This ensures that AI models and training data remain protected across their lifecycle.
The platform also introduces zero-downtime live patching for many environments, reducing maintenance interruptions for AI inference services that require continuous availability.
For enterprises handling regulated data or operating across multiple jurisdictions, VCF 9.1 provides centralized governance controls that help enforce data sovereignty and compliance policies without additional external tools.
A key aspect of VCF 9.1 is its commitment to hardware flexibility.
The platform supports GPUs from NVIDIA and AMD, alongside CPU architectures from AMD and Intel. It also integrates with high-speed networking technologies such as NVIDIA ConnectX and BlueField DPUs, enabling efficient data movement for large-scale AI workloads.
This open ecosystem approach is important because AI infrastructure is no longer defined by a single vendor stack. Instead, enterprises are increasingly mixing hardware platforms to optimize cost, performance, and availability.
By supporting multiple vendors, VCF 9.1 allows organizations to avoid vendor lock-in while still maintaining a unified operational layer.
VCF 9.1 also focuses heavily on accelerating application delivery.
The platform introduces improvements in Kubernetes scalability, faster deployment cycles, and reusable application blueprints. These features allow enterprises to move AI systems from development to production more quickly while maintaining consistency across environments.
In practical terms, this means AI teams can test models, deploy them into production, and scale them across infrastructure with fewer delays and reduced operational overhead.
The release of VMware Cloud Foundation 9.1 reflects a broader shift in the AI industry toward private cloud infrastructure for production workloads.
According to Broadcom’s Private Cloud Outlook 2026, 56% of organizations are already running or planning to run AI inference in private cloud environments, while public cloud adoption for these workloads is declining.
This shift is largely driven by concerns around cost, security, and control—three areas where VCF 9.1 directly positions itself as an alternative to public cloud infrastructure.
VMware Cloud Foundation 9.1 is not just an incremental software update. It represents a strategic attempt to redefine how enterprises build and operate AI infrastructure.
By combining Kubernetes-native architecture, multi-vendor hardware support, cost optimization, and zero-trust security, the platform aims to become a foundational layer for private AI deployment at scale.
For enterprises facing rising AI costs and increasing demand for production-grade AI systems, VCF 9.1 offers a path toward building scalable, secure, and cost-efficient AI infrastructure without relying entirely on public cloud providers.
In the long term, platforms like VCF 9.1 may play a central role in determining how and where the next generation of AI systems is actually deployed—quietly powering the infrastructure behind enterprise AI, agentic systems, and large-scale inference workloads.