How MetalSoft and VMware Cloud Foundation will deliver simpler and smarter infrastructure operations
Infrastructure teams running VMware Cloud Foundation (VCF) already have a powerful private cloud platform. But below the hypervisor — in the physical world of bare-metal servers, firmware, and hardware lifecycle — complexity has been quietly compounding. MetalSoft and Broadcom are working together to change that.
The problem nobody talks about
Private cloud platforms have come a long way. Compute, storage, and networking can be provisioned in minutes. Policies are automated. Workloads are portable. And yet, for many enterprise infrastructure teams, the experience of adding a new server to a VCF cluster still looks like this: open a manual, look up compatible firmware, run a 12-step provisioning checklist, and hope nothing breaks during the ESXi upgrade window.
VCF provides a consistent software-defined infrastructure platform across compute, storage, networking, management, and automation. But customers still face operational complexity below the VCF layer. Before an ESXi host becomes part of a VCF environment, several steps must happen correctly:
- Server discovery and inventory
- BMC access and hardware interrogation
- BIOS and firmware configuration
- Driver and firmware baseline alignment
- Network and storage readiness
- GPU, NIC, DPU, and accelerator validation
- ESXi installation and host preparation
- Hardware compliance and drift detection
- Safe decommissioning or tenant reassignment
In homogeneous environments, customers may handle these steps through OEM-specific tools and manual runbooks. But most large enterprises do not operate in homogeneous environments. They run a mix of Dell, HPE, Lenovo, Supermicro, and increasingly custom ODM infrastructure. MetalSoft publicly documents lifecycle management support across physical servers from Dell, HPE, Lenovo, and Supermicro, including firmware coordination, BIOS/BMC configuration, and hardware inventory workflows.
Beyond efficiency, there are security implications as well. In 2025, 89% YoY of attacks were AI-enabled, according to CrowdStrike. Considering this, many enterprises have set a seven-day deadline for patching vulnerabilities across their entire fleet.. In the EU, DORA Article 8 and Article 11 also require the development of automated mechanisms for firmware management, instantaneous severability, and rapid recovery.
The challenge grows even more in AI infrastructure, where server readiness is not only about CPU, memory, and storage. It includes GPU topology, high-speed NICs, DPUs, firmware, cabling, network segmentation, and multi-tenant lifecycle controls. VCF customers need these hardware operations to become more automated, repeatable, and aligned with the VCF operating model.
Hardware lifecycle management — the process of bringing physical servers into compliance, keeping firmware aligned with platform requirements, and safely retiring equipment — remains one of the last manual, error-prone frontiers in enterprise IT. It is invisible when it works and catastrophic when it doesn’t.
The cost is real. Teams report spending hundreds of hours per year on firmware research, manual pre-checks, and upgrade coordination. Hardware incompatibilities are a leading cause of unplanned downtime. And as AI infrastructure pushes server refreshes into overdrive, with GPU nodes arriving in larger quantities and requiring tighter firmware discipline, the manual approach is no longer an option.
| “The question isn’t whether to automate hardware lifecycle — it’s whether you do it before or after a production incident forces the issue.” – Infrastructure Architect, Fortune 500 Financial Services |
What the Broadcom and MetalSoft partnership enables
The MetalSoft platform is purpose-built for what VCF leaves below the hypervisor. It is an intent-based, API-first infrastructure automation engine that speaks the language of physical hardware — Redfish, IPMI, BIOS configurations, firmware catalogs, switch provisioning — and translates that into clean, auditable, policy-driven workflows.
The partnership between MetalSoft and Broadcom brings these two layers together so that the boundary between physical and virtual infrastructure becomes, for the operations team, largely invisible.
For VCF customers the goal is simple: Make physical infrastructure easier to discover, prepare, provision, validate, and operate before and after it becomes part of the VCF estate. This creates value across four major areas.
1. Heterogeneous hardware lifecycle management
Most large VCF environments include multiple hardware vendors, server generations, firmware baselines, BIOS defaults, NICs, storage controllers, and BMC implementations. That heterogeneity creates operational friction and makes in-house maintained automation nearly impossible. A lifecycle workflow that works cleanly on one server model may need a different process on another. Firmware catalog handling may vary by OEM. Redfish implementations can differ across vendors and generations. Even basic inventory and hardware state collection can require vendor-specific handling.
MetalSoft helps address this by providing a multi-vendor orchestration layer for physical infrastructure. Its MetalSoft Platform maintains distinct drivers for different vendors and sometimes different hardware or software generations, specifically to cope with different vendor implementations of similar capabilities. It also supports generic Redfish servers, which is important for broader hardware onboarding and custom infrastructure scenarios. For VCF customers, this can help standardize:
- Bare-metal server discovery
- Hardware inventory and classification
- Firmware and BIOS baseline operations
- BMC configuration
- Network and storage preparation
- ESXi host preparation workflows
- Hardware and firmware drift visibility
- Decommissioning and reuse
The advantage is not that every OEM tool disappears. Customers may still use OEM tools for deep diagnostics, warranty workflows, and vendor-specific remediation. The advantage is that VCF customers get a more consistent lifecycle operating model across heterogeneous hardware.
2. Faster infrastructure readiness for VCF clusters
VCF deployment and expansion depend on infrastructure readiness. A host is not ready simply because it is physically racked. It must have the right firmware, BIOS settings, network paths, storage visibility, ESXi image, driver alignment, and operational metadata before it can be safely introduced into the VCF environment. Manual preparation slows cluster expansion, increases upgrade risk, and makes large-scale fleet operations harder to standardize. MetalSoft’s Bare Metal Orchestrator is positioned to automate provisioning and lifecycle management across vendors, including multi-vendor firmware management and orchestration workflows. For VCF customers, this can reduce the operational gap between “hardware has arrived” and “hardware is ready for VCF consumption.” A typical flow could look like this:
- Discover new servers and hardware inventory.
- Assign servers to an infrastructure pool.
- Apply hardware baseline and BIOS/BMC policies.
- Align firmware and driver prerequisites.
- Prepare networking and storage access.
- Install or prepare ESXi.
- Validate readiness before VCF onboarding.
- Continuously monitor drift after deployment.
This is especially important for customers operating large private clouds, service-provider environments, or frequent capacity-expansion cycles.
3. AI Factory and GPU infrastructure readiness
AI infrastructure introduces a new level of hardware lifecycle complexity. An AI server is not just another compute node. It may include multiple GPUs, high-bandwidth NICs, DPUs, specialized firmware, accelerator topology, high-speed fabric dependencies, liquid-cooling considerations, and stricter workload isolation requirements. Customers building AI Factory environments need to move from manual GPU server preparation to repeatable infrastructure lifecycle automation. MetalSoft positions its AI Factory solution as a bare-metal and network orchestration platform for enterprise AI factories and GPU-as-a-Service providers. Its AI Factory materials call out automated discovery of servers, switches, DPUs, cabling, firmware, and GPUs across multi-vendor environments. For VCF customers, this is highly relevant. VCF can provide the enterprise cloud operating model for AI infrastructure, while MetalSoft helps automate the underlying physical lifecycle required to make GPU infrastructure ready, consistent, and reusable. This can help customers with:
- GPU server discovery and classification
- AI cluster buildout readiness
- Firmware and driver baseline consistency
- Network segmentation for AI workloads
- GPU-as-a-Service infrastructure preparation
- Tenant deprovisioning and infrastructure reset
- Multi-vendor GPU fleet lifecycle
- Faster AI capacity onboarding
- Firmware drift detection
MetalSoft also describes secure tenant deprovisioning capabilities for AI factory and GPUaaS environments, including storage sanitization, GPU persistence reset, and removal of network segments to prevent tenant data leakage. That matters because AI infrastructure is increasingly multi-tenant. Enterprise platform teams, research teams, business units, and service providers all need isolated access to expensive GPU capacity. Operational safety and lifecycle reset become as important as initial provisioning.
4. A practical path for custom ODM infrastructure
Customers including CSPs, sovereign cloud providers, AI infrastructure builders, and very large enterprises are increasingly evaluating ODM or custom hardware options to optimize cost, supply chain, power, density, or AI server design. MetalSoft’s use of vendor-specific drivers, generic Redfish support, and multi-vendor infrastructure orchestration creates a practical path to onboard a broader set of hardware platforms. This does not mean every custom ODM platform becomes automatically supported. ODM readiness still depends on hardware validation, Redfish/BMC quality, firmware integration, support agreements, and the customer’s operational requirements. But the partnership creates a better foundation for:
- Custom hardware onboarding
- ODM server discovery
- Redfish-based lifecycle workflows
- Hardware baseline automation
- Repeatable ESXi preparation
- Validation of non-traditional server platforms
- Reduced dependence on one-off scripts and manual runbooks
For customers pursuing AI, sovereign cloud, or cost-optimized private cloud, this flexibility is strategically important.
How VCF and MetalSoft fit together
The partnership is best understood as a layered operating model.
+———————————————————————————–+
| VMware Cloud Foundation (Private Cloud) |
| [ Fleet Manager ] [ vCenter ] [ vSAN ] [ NSX Networking ] |
+———————————————————————————–+
Standardized APIs & Declarative Policies
+———————————————————————————–+
| MetalSoft Bare-Metal Orchestration Engine |
| [ Multi-Vendor Drivers ] [ Redfish Engine ] [ Switch Fabric ] |
+———————————————————————————–+
Direct Hardware Control via BMC, Redfish & IPMI
+———————————————————————————–+
| Heterogeneous Physical Infrastructure |
| [ Dell ] [ HPE ] [ Lenovo ] [ Supermicro ] [ ODM ] |
+———————————————————————————–+
Together, VCF and MetalSoft can help customers bridge the gap between physical infrastructure and cloud operations. The value is not simply “bare-metal provisioning.” The value is a more complete Day 0 to Day N lifecycle model across the infrastructure stack.
What this means for enterprise platform teams
Enterprise platform teams are being asked to deliver more infrastructure, faster, with stronger governance and less operational risk. They need to support traditional enterprise applications, cloud-native platforms, AI workloads, GPU clusters, and sometimes bare-metal consumers — all while managing heterogeneous hardware. The Broadcom and MetalSoft partnership helps shift bare-metal operations from manual coordination to platform-driven automation. Instead of treating hardware readiness as a separate pre-VCF process, customers can create a more connected lifecycle: Discover physical infrastructure, prepare it through policy-driven workflows, validate it for VCF use, onboard it into the VCF environment, and continue monitoring compliance over time. That is the foundation for a more complete private-cloud operating model.
Partnership value at-a-glance
| Value Driver | What It Means for You |
| Faster server commissioning | Declarative, policy-driven provisioning replaces multi-step manual checklists. New servers join VCF clusters in a validated, compliant state. |
| Faster zero-day vulnerability response | Lower security risk, faster reaction to incidents, easier compliance with DORA and similar legislation. |
| Elimination of HCL drift | Every server in the fleet carries firmware aligned to the running VCF platform version. HCL mismatches are detected and resolved before they create support exposure. |
| Coordinated upgrade operations | Firmware and platform lifecycle events are planned together, reducing maintenance window length and workload disruption. |
| Secure, auditable decommissioning | Standards-compliant data wipe with cryptographic attestation eliminates manual steps and creates defensible audit trails. |
| Multi-vendor fleet simplicity | A single management model across mixed OEM /ODM environments reduces the operational cost of vendor diversity. |
| AI infrastructure readiness | GPU server provisioning, accelerator firmware management, and AI workload-specific validation extend the same lifecycle discipline to the highest-stakes part of the modern data center. |
| Infrastructure as code for hardware | Server configuration is version-controlled, repeatable, and auditable, bringing hardware lifecycle into the same operational model as software delivery. |
Different kind of infrastructure partnership
Most infrastructure partnerships are integrations; one product calls another product’s API and the result is a checkbox on a compatibility matrix. This partnership is different in intent.
VMware Cloud Foundation handles the virtualization layer: compute, storage, networking, and the policy-driven management that makes private cloud operationally comparable to public cloud. MetalSoft handles the physical layer: the servers, the firmware, the hardware lifecycle events that happen before VCF can even see the resource.
These two layers are complementary by design, not competitive. And the teams working on this partnership are focused on a specific outcome: making the experience of running infrastructure on VCF simpler and more reliable at the hardware level, without requiring customers to change how they use VCF.
The goal is for an infrastructure team to be able to add 50 servers to a VCF cluster — or decommission 50 servers, or upgrade the firmware on 200 servers — without it being the most stressful week of the quarter.
| Private cloud should be as operationally simple as public cloud. This sentiment is true for workloads. This partnership is about making it true for the hardware those workloads run on. |
Discover more from VMware Cloud Foundation (VCF) Blog
Subscribe to get the latest posts sent to your email.
