The most costly IBM licensing mistakes.
Not all licensing mistakes cost the same. A handful of them dwarf the rest, because they each switch a count from the workload to the whole environment. The expensive errors are not the small miscounts. They are the structural ones that let IBM default to the largest possible base, and every one of them is reversible when the position is rebuilt.
Letting ILMT lapse under sub-capacity
The single most expensive mistake is running sub-capacity eligible software without the inventory evidence to support it. Sub-capacity licensing lets you license only the cores running IBM software instead of the whole physical host, but it requires ILMT or an approved tool deployed within 90 days of first eligible deployment, running continuously, with quarterly reports retained two years. Miss any of those and IBM defaults to full-capacity charging. A product using four of thirty two host cores can flip from a few hundred PVU to several thousand. That is the same software, counted against the whole box, purely because the evidence was missing.
Ignoring the container and Cloud Pak cluster rule
For container and Cloud Pak workloads, non-compliance does not default to one host. It charges every core in the cluster. A modest overrun on part of a large OpenShift cluster can be counted against the entire cluster, producing a finding many times the size of the workload that caused it. This is the modern equivalent of the full-capacity default, and it is becoming the costliest mistake as more of the IBM portfolio moves onto containers and the VPC metric.
Using bundled software beyond its scope
IBM products bundle components, such as a Db2 included with another product, under restricted use rights. Using a bundled component for a workload outside the bundle's allowed scope looks, to discovery, like an unlicensed standalone product. Because the bundled component is everywhere, the resulting finding can be large and widespread. Mapping each instance to how it was entitled, and separating restricted use bundles from full entitlements, is what keeps legitimate bundled use from being counted as standalone licenses.
Dropping support without expecting the audit
Letting subscription and support lapse to save cost is one of the most reliable audit triggers there is. The saving is immediate and visible, the audit that follows is neither, and it arrives after deployment has drifted away from entitlement. The mistake is not the decision itself but making it blind, without first rebuilding the license position so you know what an audit would find. The cost is the gap between what you thought you were running and what you actually were, surfaced at the worst possible moment.
Never establishing a baseline
Underneath all of these sits one root mistake: never building an independent picture of what you deploy against what you own. Without a baseline, every other error compounds unseen, entitlements you hold go uncredited, and the first time anyone reconciles the estate is when IBM does it for you. A maintained baseline turns each of the expensive mistakes above into something you catch early rather than discover under audit.
The costliest IBM licensing mistakes share a pattern: each lets the count default to the largest base, whether that is the full host, the whole cluster, or an unlicensed reading of bundled use. Lapsed ILMT, the container cluster rule, bundled misuse, dropping support blind, and the absence of a baseline are the ones that move real money. All are reversible when the position is rebuilt from clean evidence and the findings are challenged line by line.
Has a structural mistake inflated your exposure?
Our Audit Negotiation engagement challenges the findings that rest on these mistakes line by line, reverses the full-capacity and cluster defaults where the evidence supports it, and credits the entitlements an audit left out.
See Audit Negotiation →The IBM Audit Brief
Audit triggers, ILMT pitfalls, and settlement tactics for IBM software buyers.
Independent, buyer side IBM software audit defense and negotiation. Not affiliated with IBM Corporation.