Key Points
- VCF 9.1 supports mixed compute infrastructure across AMD, Intel and Nvidia hardware
- Broadcom survey shows 56 per cent of organisations prefer private cloud for AI inference
- Platform claims up to 40 per cent reduction in server costs through memory tiering
Broadcom has announced VMware Cloud Foundation (VCF) 9.1, a private cloud platform designed for enterprises running artificial intelligence workloads on their own infrastructure rather than public cloud services. The platform supports hardware from AMD, Intel and Nvidia, allowing organisations to mix different processors and graphics cards in the same deployment.
VCF is a software suite that bundles virtualisation, storage and networking tools to let companies run multiple applications on shared physical servers. The 9.1 release adds features specifically for AI applications, including inference — the process of running trained AI models to generate outputs — and what Broadcom calls agentic AI, referring to AI systems that can take actions autonomously.
A preview of Broadcom’s Private Cloud Outlook 2026 report, cited in the announcement, indicates that 56 per cent of organisations surveyed are running or planning to run production AI inference on private cloud infrastructure. Public cloud usage for the same purpose stood at 41 per cent, down 15 percentage points year-on-year, according to the company.
The survey also found that 62 per cent of IT leaders reported being very or extremely concerned about generative AI infrastructure costs. Separately, 36 per cent said AI is creating new requirements for data protection, privacy controls and risk management, Broadcom stated.
“As more enterprises turn to AI for driving competitive advantage, they face three critical challenges: data and IP privacy concerns, surging infrastructure costs, and their readiness for the world of agentic AI,” said Krish Prasad, senior vice president and general manager, VMware Cloud Foundation Division, Broadcom.
Cost reduction claims
Broadcom claims VCF 9.1 can deliver up to 40 per cent reduction in server costs through what it calls intelligent memory tiering — a technique that automatically moves data between faster and slower memory based on how frequently it is accessed. This applies to clusters running a mix of AI and non-AI workloads, the company stated.
The platform also claims up to 39 per cent lower total cost of ownership for storage through enhanced compression and deduplication for AI data pipelines, according to Broadcom’s announcement. Deduplication removes redundant copies of data to reduce storage requirements.
For Kubernetes — an open-source system for managing containerised applications — Broadcom claims up to 46 per cent reduction in operational costs for running AI workloads at scale. The company also stated the platform delivers four times faster cluster upgrades and doubled fleet management capacity to 5,000 hosts.
Advertisement
Enterprise deployment features
VCF 9.1 includes what Broadcom describes as multi-tenant infrastructure for AI isolation. This allows enterprises and service providers to run multiple AI projects on shared hardware while maintaining security boundaries between them.
The feature is relevant for Indian enterprises subject to data localisation requirements under regulations such as the Digital Personal Data Protection Act.
The platform supports automated operations across distributed and air-gapped environments — networks that are physically isolated from the internet for security reasons. This capability addresses deployment scenarios in regulated sectors such as banking and defence where systems cannot connect to external networks.
Your Questions, Answered
What is VMware Cloud Foundation 9.1?
VCF 9.1 is a private cloud software platform from Broadcom that bundles virtualisation, storage and networking tools. The 9.1 release adds features for running AI workloads on enterprise-owned infrastructure.
Which hardware does VCF 9.1 support?
The platform supports mixed compute infrastructure across AMD, Intel and Nvidia processors and graphics cards, allowing enterprises to choose hardware from multiple vendors.
What cost savings does Broadcom claim for VCF 9.1?
Broadcom claims up to 40 per cent reduction in server costs through memory tiering, 39 per cent lower storage costs through compression and deduplication, and 46 per cent reduction in Kubernetes operational costs.
Why are enterprises preferring private cloud for AI?
According to Broadcom’s survey, 62 per cent of IT leaders are concerned about generative AI infrastructure costs, while 36 per cent cite data protection, privacy and security requirements driving private cloud adoption.







