Broadcom Update to VCF Platform Promises to Reduce IT Costs

Broadcom today updated its VMware portfolio to reduce the cost of running workloads while at the same time adding support for graphics processing units (GPUs) from AMD and networking infrastructure from Arista Networks.

Prashanth Shenoy, chief marketing officer and vice president of marketing for cloud platform, infrastructure, and solutions for the VCF division of Broadcom, said version 9.1 of the VMware Cloud Foundation (VCF) makes it less expensive to run a variety of workloads while at the same time serving to make IT environments more resilient.

For example, a tiering capability has been improved to make better use of solid-state drives (SSDs) connected to a Non-Volatile Memory Express (NVMe) backplane as a second tier of memory to reduce the cost of running various classes of workloads by as much as 40%.

Broadcom has also expanded the number of hosts that VCF can effectively manage to 5,000 while at the same time improving by a factor of four the rate at which individual clusters can be upgraded. For example, Broadcom claims the latest version of VCF provides a 2.6x increase in the number of clusters that can run compared to the previous release, while reducing the time to deploy and upgrade by 70% and 75%, respectively.

Additionally, Broadcom claims to have reduced by up to 39% the total cost of ownership of storage for artificial intelligence (AI) workloads via support for enhanced compression and deduplication capabilities.

In addition to previewing support for object-based storage, Broadcom is also now providing access to live application stack blueprints that capture workloads spanning multiple virtual machines to provide IT teams with a set of reusable templates for rapidly deploying applications.

Broadcom has also added support for integrated container management, zero-trust segmentation extensions to Kubernetes clusters, sovereign recovery and continuous patching for AI workloads, centralized monitoring and automated desired state remediation for workloads and VCF stack components and live patching.

On the security front, VCF 9.1 now provides support for centralized tagging, pre-defined security profiles, delegated firewall configurations and ingress web application security along with on-premises ransomware recovery capabilities and support for CrowdStrike Falcon Endpoint Security tools.

Finally, Broadcom has updated VMware Avi Load Balancer and VMware vDefend to eliminate hardware appliance requirements for AI inference endpoints and agentic applications, while adding support for Kubernetes clusters and the Model Context Protocol (MCP).

Overall, Broadcom reports that more than 2,000 IT organizations have now deployed VCF 9.0 in the last year, spanning more than 19 million processors.

In general, many of those organizations are adopting VCF to deploy private clouds to run AI workloads, noted Shenoy. A preview of Broadcom’s Private Cloud Outlook 2026 survey finds more than half of organizations surveyed (56%) are running or planning to run production inferencing in a private cloud. Reliance on public cloud services for AI inference workloads was 41%. A full 62% of survey respondents are also either very or extremely concerned about generative AI infrastructure costs while 36% report AI is driving new requirements for data protection, privacy, security controls and risk management.

Broadcom, of course, has been making a case for delivering more value at a higher cost since acquiring VMware. As such, the latest update to VCF is part of a continuing theme that seems to be resonating with enough organizations to provide a return on a $61 billion acquisition that was completed nearly four years ago.