SAP AI Launchpad
As of 2026-11-07
What is SAP AI Launchpad?
SAP AI Launchpad is a multitenant SaaS management layer on SAP BTP for AI use cases across multiple AI runtime instances, including SAP AI Core — it does not include the runtimes themselves. Its Generative AI Hub app (prompts, deployments, orchestration, monitoring/comparison views) requires SAP AI Core's Extended service plan; the Standard plan does not unlock it, and there is no downgrade path from Extended.
A management layer, not a runtime
SAP's own documentation is precise about what SAP AI Launchpad is: a multitenant software-as-a-service application on SAP Business AI Platform that lets customers and partners manage AI use cases across multiple AI runtime instances — including, but not limited to, SAP AI Core. The AI runtimes themselves are not included in the Launchpad subscription; Launchpad is the cockpit, not the engine. That distinction matters for scoping: a client asking "do we need AI Launchpad" is really asking whether they need a management layer across one or several SAP AI Core tenants, not whether they need compute. SAP AI Launchpad runs exclusively in the SAP BTP Cloud Foundry environment.
Six documented capabilities
SAP groups what AI Launchpad does into six areas: connecting to and switching between multiple AI runtime instances; interacting with SAP AI Core resource groups and their assets (models, executions, deployments); managing the generative AI lifecycle — prompts, experimentation, administration — through the Generative AI Hub app inside Launchpad; managing the AI use-case lifecycle from training through deployment; tracking consumption statistics and resource utilization; and handling authentication and operational administration for the connected AI runtimes. Resource groups are managed from Launchpad itself: create, edit and delete a resource group (a separate flow exists specifically for a resource group intended for AI data management), then use it to scope models, executions and deployments the same way you would from SAP AI Core's own API.
Why it matters
- AI Launchpad manages runtimes, it does not provide compute itself — scoping 'do we need Launchpad' is a different question from 'do we need AI Core capacity,' and confusing the two derails a proposal.
- The Generative AI Hub inside Launchpad is licensing-gated on SAP AI Core's Extended service plan, not a Launchpad setting — a client on the Standard plan will not see it no matter how Launchpad is configured.
- The upgrade from Standard to Extended is one-way: data and models carry over, but there is no documented downgrade path, which changes how conservatively a plan decision should be made.
Key points
- AI Launchpad is a multitenant SaaS management layer on SAP BTP, running exclusively in the Cloud Foundry environment; AI runtimes (e.g. SAP AI Core) are not part of the subscription.
- Six documented capabilities: multi-runtime connection, resource-group interaction, Generative AI Hub lifecycle management, use-case lifecycle management, consumption/utilization analytics, and runtime administration.
- Resource groups are created, edited and deleted directly from Launchpad, including a dedicated flow for AI-data-management resource groups.
- Generative AI Hub access requires SAP AI Core's Extended service plan (Standard does not include it); upgrade Standard→Extended is supported, downgrade is not; a 30-day free trial exists for exploration.
- Documented flows include creating deployments (including specifically for the orchestration service), creating/deleting prompts, prompt optimization, and monitoring via compare-executions, compare-models, compare-runs and compare-model-metrics.
- Two operating roles frame who uses Launchpad: an AI scenario producer develops and productizes AI use cases on one or more runtimes; an AI scenario consumer subscribes to and uses an already-productized scenario.
- SAP AI Core enforces hard tenant-wide quotas Launchpad surfaces indirectly: a maximum of 50 resource groups and 5 tenant-wide generic secrets per tenant — plan scenario count against this before scaling.
- Extended plan: pay per resource plus model token usage, converted to GenAI tokens and billed in BTP capacity units (SAP Note 3437766); Standard is resource-based only.
Terms used on this page
- AI runtime
- An execution engine such as SAP AI Core that SAP AI Launchpad connects to and manages, but does not itself include in the Launchpad subscription.
- Extended service plan
- The SAP AI Core production plan that includes Generative AI Hub access, as distinct from the Standard plan, which does not.
- Resource group (AI Launchpad view)
- The same SAP AI Core isolation unit for models, executions and deployments, created/edited/deleted from within Launchpad.
- Prompt optimization
- A documented AI Launchpad flow for refining a saved prompt inside the Generative AI Hub's prompt registry.
- Compare executions/models/runs/metrics
- AI Launchpad's monitoring views for evaluating generative and predictive AI work side by side rather than reading separate outputs.
Sources
- GitHub SAP-docs — sap-artificial-intelligence: What Is SAP AI Launchpad? (sap-ai-launchpad/what-is-sap-ai-launchpad)
- GitHub SAP-docs — sap-artificial-intelligence: SAP AI Core Service Plans (Standard vs Extended)
- SAP Help Portal — What Is SAP AI Launchpad?
- SAP Help Portal — Create a Resource Group | SAP AI Launchpad
- GitHub SAP-docs — Using SAP AI Launchpad (AI scenario producer/consumer roles, SAP-authored mirror)
- GitHub SAP-docs — Generative AI Hub inside SAP AI Launchpad (SAP-authored mirror)
- GitHub SAP-docs — AI Scenario Producer (SAP-authored mirror)
- GitHub SAP-docs — AI Scenario Consumer (SAP-authored mirror)
- GitHub SAP-docs — SAP AI Launchpad concepts (SAP-authored mirror)
- SAP Discovery Center — SAP AI Core service catalog entry (plans, pricing entry point)
- SAP Help Portal — Prompt Registry, generative AI hub (Git-based sync API for CI/CD, 2026)
- SAP Help Portal — Metering and pricing for generative AI on SAP AI Core (AI Units mechanics behind the Extended plan)
Full card with the Consultant plan — €19.90/month. What the full card adds: the full decision framework · the SAP vs Snowflake / Databricks / Fabric comparison · the common pitfalls and their fix · the cheat sheet · the architecture schemas · the code blocks · the facts worth quoting.
SAP AI Launchpad architect: what is decided before the first click?
An SAP AI Launchpad architect decides the structure the console will display. Resource groups come first: up to 50 per tenant by default, no moving of runtime objects between them afterwards, and the finest cost grain the platform reports natively — so one per use case and environment. Role collections come second: experimenter, manager and administrator rights belong to different people. Evidence is third: inference observability records model, tokens and latency only for requests that send its headers. And Launchpad stays out of the run path: it is a console, not a runtime.
What should an SAP AI Launchpad freelancer be able to do?
An SAP AI Launchpad freelancer is, in practice, an AI Core or generative AI hub consultant who works through the console. In the contract corpus of 10 October 2026 no posting names Launchpad as a skill and no day rate is published, so ask for three jobs: testing whether a model can summarise or classify business text before code is written, setting up a Grounding Management repository for an agent, and reading consumption and evaluation evidence for a governance review.
Related guides
These guides cite this page as a source.
- How to Become SAP AI Consultant: 90-Day Plan from BW/SAC
- Freelance SAP AI Architect: What to Hire For
- Freelance SAP AI consultant: scope, formats, checks
- SAP AI Architect: Layers, Grounding, Agents, EU AI Act
- SAP AI Core Consultant: Scope, Skills and Deliverables
- SAP Generative AI Hub Consultant: Scope and Deliverables
Read next
Pages this one leads to, each answering a question in full:
Open missions
102 live roles and contracts in SAP AI & Joule right now — see the SAP AI & Joule missions · freelance SAP contracts.