Lesson 1 of 7
SAP AI Core, SAP AI Launchpad and the generative AI hub
8 min
Listen to this lesson
Theo, Contoso's integration lead, wants to build a small application: when an inbound IDoc fails, key users get a plain-language explanation and a suggested owner, grounded in Contoso's own interface runbooks, with customer data masked before anything reaches a model. Joule doesn't offer this out of the box, so it becomes a custom application on Contoso's SAP BTP subaccounts using the generative AI hub in SAP AI Core. We'll call it the CMC IDoc Explainer and build its design through this course. Facts are as of 17 October 2026; the hub changes monthly and SAP TechEd (27 to 29 October 2026) may add more.
Three names, one stack
SAP's documentation is precise about the relationship. SAP AI Core "provides the runtime and lifecycle management foundation for AI scenarios" on SAP BTP: deployments, executions, resource groups and security. SAP AI Launchpad is the user interface for managing AI scenarios and includes the generative AI hub screens: a prompt editor, chat and an orchestration workflow builder. The generative AI hub is "a capability of SAP AI Core and SAP AI Launchpad" that gives governed access to large language models and the orchestration features around them. In SAP's words, it "doesn't replace SAP AI Core; it extends it."
Service plans
Which plan you choose decides whether the hub is available at all. SAP's service plans page: use Standard "for production workloads without generative AI" and Extended "for production workloads with generative AI hub access". Standard is priced per resource plus a baseline charge; Extended per resource plus model and token usage. You can upgrade from Standard to Extended but not downgrade. For learning, SAP offers a 30-day basic trial of the generative AI hub. SAP's metering page states it directly: "The generative AI hub is available only as part of the extended service plan."
| Plan | Generative AI hub | Pricing basis | Use for |
|---|---|---|---|
| Standard | Not included | Per resource plus baseline charge | Classic ML training and serving |
| Extended | Included | Per resource plus model and token usage | Production generative AI |
| 30-day basic trial | Included for exploration | Free, time-limited | Learning and prototypes, never production data |
SAP AI Core service plans as documented on SAP Help.
Resource groups and scenarios
Everything in SAP AI Core lives in a resource group, and every API call names one in the AI-Resource-Group header. Resource groups separate teams or use cases, carry their own deployments and secrets, and are limited to 50 per tenant. Generative models are reached through two global scenarios managed by SAP AI Core: foundation-models, where you deploy a specific model, and orchestration, the harmonized pipeline covered in Lesson 3. SAP's documentation usually provides an orchestration deployment in the default resource group; you create one in another resource group when you need to separate use cases.
Rules of use
SAP's documentation includes a usage restriction worth knowing: SAP AI Core, including the generative AI hub, "must not be used to generate synthetic data for training or fine-tuning models", unless an exception is explicitly approved and documented. It also documents which features are not offered for sovereign cloud landscapes. Read the service's terms with your architect before a design depends on a feature.
What would you do?
Theo's prototype works in the 30-day trial. To save time, a colleague suggests creating the production SAP AI Core instance in cmc-ext-prod on the Standard plan, because "it's cheaper and we can switch later". What should Theo say?
- SAP AI Core is the runtime; SAP AI Launchpad is the UI; the generative AI hub is a capability of both.
- Generative AI hub needs the extended service plan; a 30-day basic trial exists for learning.
- Resource groups separate use cases (max 50 per tenant); calls name one in the AI-Resource-Group header.
- Models are reached via the foundation-models or orchestration scenario.
Check your understanding · not graded
Lesson check
1. How does SAP describe the generative AI hub's relationship to SAP AI Core?
2. Which SAP AI Core service plan includes the generative AI hub?
3. Which header names the resource group in an SAP AI Core API call?
4. What does SAP's documentation prohibit using SAP AI Core, including the hub, for without an approved exception?