watsonx Licensing Primer.

The watsonx licensing reference for the buyer side. The Resource Unit metric, the foundation model class pricing, the watsonx.ai inference and training entitlement, the watsonx.data lakehouse capacity model, the watsonx.governance seat structure, and the Cloud Pak for Data on premises path.

What is inside.

• The three product lines (watsonx.ai, watsonx.data, watsonx.governance) and the boundary between each. The integration to the Cloud Pak for Data on premises path. The SaaS versus self managed decision.
• The Resource Unit metric. The conversion from RUs to consumption (inference calls, training jobs, governance seats, lakehouse capacity). The minimum commitments and the variable consumption profile.
• Foundation model class pricing tiers. Small, medium, large, frontier. The Granite family pricing relative to third party models on watsonx. The deprecation and substitution policy.
• watsonx.data capacity sizing. The compute pool versus storage capacity model. The integration to the existing Db2 and Cognos estate. The cross product entitlement consideration.
• watsonx.governance seat structure. The model risk team seats, the platform seat, the model registry capacity. The integration to the regulated AI workload programme.
• The Cloud Pak for Data on premises path for watsonx. The bundled OpenShift entitlement, the VPC sizing, and the data sovereignty case for self hosting.
• The watsonx multi year commitment frame. The ramp curve, the true up and true down provisions, the price protection, and the deprecation handling.
• The buyer side close position on watsonx. The four commercial asks, the three contractual asks, the two operational disciplines. The integration to the broader IBM relationship.

Who this is for.

• The chief data officer setting the enterprise AI commercial frame.
• The procurement lead negotiating the first watsonx commitment.
• The platform architect operating the watsonx stack.
• The CFO evaluating the multi year AI run rate.