VMware License Harvesting Strategies

License harvesting is one of the most under-utilised tactics in the VMware cost-optimisation toolkit. In simple terms, harvesting means systematically identifying licences in your estate that are no longer producing value — through over-provisioning, decommissioned workloads, acquired entitlements that overlap with existing ones, or mis-allocated SKUs — and either re-deploying them, retiring them, or using them as negotiating leverage with Broadcom.

Under the old VMware perpetual model, harvesting was largely a housekeeping exercise. Under Broadcom's subscription model it is a strategic discipline that can shave 15-40% off your annual VMware spend without changing a single workload. This article walks through the harvesting strategies that produce the highest return, and the specific traps to avoid as Broadcom continues to evolve its audit tooling.

Why harvesting matters under Broadcom

The Broadcom subscription model charges per core, per year, and bundles previously independent products (vSphere, vSAN, NSX, Aria) into VCF SKUs. The model has two consequences that make harvesting essential. First, every idle core is a recurring cost — not a sunk capex. Second, oversized clusters and over-provisioned hosts are directly visible in Broadcom's audit telemetry, which means that idle capacity is also audit exposure.

Mature harvesting programmes target three categories of waste: unused entitlements (capacity you bought but never deployed), under-utilised entitlements (capacity deployed but running below threshold), and over-licensed deployments (workloads that have been assigned more entitlement than they consume). Each category requires a different approach.

Strategy 1 — The decommissioning audit

Every IT estate has zombie workloads: VMs that were stood up for a project five years ago, never decommissioned, consuming cores and host resources for no business reason. Under perpetual licensing, those VMs were free to ignore. Under subscription, every one of them is a line item on next year's renewal.

A structured decommissioning audit typically uncovers 8-15% of an estate's running workloads as candidates for retirement. The methodology is straightforward: pull a 90-day utilisation report from Aria Operations or vCenter, filter for VMs with sustained CPU below 5%, network I/O below 1 MB/min, and no recent administrative activity. Each candidate is then verified with the workload owner before retirement.

The harvested capacity can either reduce the next renewal commitment or be re-allocated to growth workloads that would otherwise have required new entitlements. Either way, the cost saving is real and recurring.

Strategy 2 — Cluster right-sizing

Many VMware clusters were sized for peak load conditions that never materialise. A cluster designed for failure tolerance N+2 with 20% headroom often runs at 35-40% utilisation. Under Broadcom's per-core subscription, those unused cores cost money every year.

Right-sizing analysis evaluates each cluster against its actual workload pattern, not its design pattern. Tools like Aria Operations capacity planning, or third-party platforms like CloudHealth or Densify, can identify clusters that could be consolidated, hosts that could be retired, or DRS configurations that could be tightened. The harvested cores convert directly into reduced subscription commitment.

Watch out for cluster boundaries

Broadcom licenses at the host level, and cores are counted per physical CPU regardless of how DRS distributes workloads. Consolidating two 16-core hosts into one 32-core host does not reduce your core count. The harvesting opportunity lies in retiring hosts entirely, not in reshuffling capacity within the same hardware envelope.

Strategy 3 — Acquisition reconciliation

If your organisation has acquired companies in the last five years, you almost certainly have orphaned VMware entitlements buried in those acquired estates. The acquired company's licences may still sit in their original VMware account, may overlap with your existing entitlement, or may include perpetual licences with paid support that you can leverage commercially.

Reconciling those entitlements requires patience: every acquired company's licence ledger needs to be pulled, every ELA amendment reviewed, every transferred licence documented. The work is tedious but the payoff is significant — we routinely see acquired entitlements worth $200K-$2M sitting unused because nobody traced them through the integration.

Strategy 4 — SKU rationalisation

Pre-Broadcom VMware customers often hold a chaotic mix of SKUs accumulated over years: Standard, Enterprise Plus, vCloud Suite, vRealize Suite, vSAN editions of various flavours. Many of those SKUs overlap functionally, and many include components that are no longer used.

SKU rationalisation maps every SKU to actual feature usage in your environment. The output is a target SKU portfolio that delivers the same functional coverage with fewer line items, lower complexity, and — typically — lower aggregate cost. Under Broadcom's VCF bundling, SKU rationalisation also informs the negotiation: if you can demonstrate that you need only vSphere Foundation rather than full VCF, you have leverage against Broadcom's default position of "everyone needs VCF."

Strategy 5 — Pre-audit harvesting

If you suspect a Broadcom audit is coming, the worst time to harvest is during the audit. The best time is in the 90-180 days before notification. Pre-audit harvesting reduces your visible deployment footprint, eliminates over-provisioning that would otherwise create false compliance signals, and gives you clean, defendable utilisation data to present.

This kind of pre-emptive harvesting is exactly where independent audit defence advisors add value. — the firm we recommend most often for Broadcom audit defence — builds harvesting into their pre-audit preparation work. They will inventory your estate, identify harvest candidates, validate decommissioning safety, and produce the documentation Broadcom auditors will eventually request, but in a form that supports your position rather than theirs.

The traps to avoid

Harvesting has a few well-known pitfalls. The first is decommissioning without proper service-owner sign-off — the workload you retire on Tuesday may be the one that processes quarterly billing on Friday. Always validate before retiring.

The second trap is harvesting that breaks high-availability commitments. Reducing host counts without recalculating HA admission control can leave clusters with insufficient failover capacity, which is a real operational risk regardless of licensing savings.

The third trap is harvesting that triggers SKU re-tiering. Some VMware bundles include minimum-cluster-size or minimum-core-count thresholds. Harvesting below those thresholds can force a SKU downgrade that loses features you actually use. Map every harvest target against the SKU rules before retiring capacity.

Building harvesting into the operating rhythm

The most effective VMware customers run harvesting on a continuous cycle, not as a one-off project. Quarterly utilisation reviews, annual SKU rationalisations, and event-driven harvesting (every acquisition, every divestiture, every major application retirement) keep the estate lean and the audit exposure low.

The cultural shift required is small but important: harvested capacity is not "spare" capacity to be re-deployed casually. It is recovered budget that goes back to the IT cost base or fuels strategic investment. Treating harvested capacity as cash, not as headroom, is what separates programmes that produce real savings from those that just produce reports.

Quantifying the opportunity

For a typical mid-market VMware customer with 200-500 hosts, structured harvesting produces savings in the $300K-$1.5M annual range. For large enterprise customers with thousands of hosts, the figure routinely exceeds $5M annually, with peak engagements producing eight-figure savings against initial Broadcom renewal positions.

The investment to capture those savings is modest: a structured harvesting programme typically requires 4-8 weeks of dedicated analysis and 1-2 FTEs of operational follow-through. The ROI is among the highest available in software asset management.

Final thought

Under Broadcom, every idle core is a tax on your IT budget that compounds annually. Harvesting is the most direct, lowest-risk way to reduce that tax.

Frequently asked questions

How long does a typical harvesting programme take?

For a mid-size enterprise (200-500 hosts), a structured harvesting programme typically runs 4-8 weeks of analysis followed by 60-90 days of operational follow-through to actually retire identified candidates. Larger enterprises can take 3-6 months for a full first-cycle programme; subsequent cycles are faster because the discovery and tooling work is already complete. The longest pole in the schedule is usually the workload-owner validation work — getting business stakeholders to confirm that identified zombie workloads really can be retired.

What tools work best for harvesting analysis?

VMware Aria Operations is the natural starting point because it sits directly on the vCenter inventory and provides capacity, performance, and configuration data in one place. Third-party platforms like Densify, CloudHealth, and Turbonomic add capacity planning and right-sizing recommendations that go beyond what Aria provides natively. For acquisition reconciliation work, traditional software asset management tools (Flexera, Snow, ServiceNow SAM) help map purchase records to deployed entitlement. The right toolset depends on what the enterprise already has and what gaps need to be filled.

Will harvesting trigger a Broadcom audit?

Harvesting itself does not trigger audits — the reverse is closer to true. Visible over-provisioning, sprawled deployments, and unexplained capacity drift are what trigger Broadcom audit attention. Harvesting reduces that surface area and produces clean, defendable utilisation data. The risk is in harvesting clumsily: dropping commitments mid-contract without contractual basis, or retiring workloads that turn out to be essential. Proper harvesting governance avoids both risks.

How does harvesting interact with VCF subscription commitments?

VCF subscriptions typically commit to a minimum core count for a fixed term. Harvesting can reduce deployed capacity below the commitment, but the contractual commitment still applies. The most effective harvesting programmes time their major reductions to coincide with renewal windows, where the harvested capacity translates directly into a reduced new commitment. Mid-term harvesting still has operational value (smaller estate to manage, lower audit exposure) but the financial benefit is deferred to renewal.

Can harvesting work for very small VMware estates?

Yes, but the ROI curve is different. Very small estates (under 50 hosts) often see the largest relative savings from harvesting, because over-provisioning ratios in small environments are typically higher than in large ones. The absolute savings are smaller, but on the percentage basis the impact is real. For smaller estates, the work is also faster — typically completable in 2-4 weeks rather than 2-4 months.