Sunil Singh

Senior Manager — SAP Analytics Consulting, Deloitte Consulting GmbH

March 28, 2026 • Updated November 2026

13 min read

Flagship Editorial Architecture + Strategy Beginner to Expert

How SAC, Datasphere, and Business Data Cloud fit together

The most confusing part of SAP's analytics platform is not what each product does—it's how they relate. SAP Analytics Cloud handles consumption and planning. Datasphere manages semantic data fabric and governance. Business Data Cloud unifies the platform as a commercial bundle, adding data products, intelligent content, and BDC Connect for zero-copy sharing, while SAP Integration Suite on SAP BTP connects the sources. This article explains the architecture, the data flow, and how to think about building on each layer.

The Three-Layer Analytics Architecture

The SAP analytics platform organizes into three clear layers, each with distinct purpose and audience:

Consumption & Planning (top): Where business users explore, decide, and act. SAC provides the primary interface.
Semantic & Modeling (middle): Where architects and data teams define trusted business meaning. Datasphere is the central hub.
Platform & Integration (bottom): Where Business Data Cloud bundles SAC, Datasphere, SAP HANA Cloud, SAP Master Data Governance, SAP BW (private cloud edition), and the partner engines SAP Databricks and SAP Snowflake into a unified commercial offering with capacity pricing, while SAP Integration Suite on SAP BTP connects the sources.

The key architectural insight: SAC and Datasphere were never designed to be chosen as either-or alternatives. They are designed to work together, with Datasphere feeding semantic context into SAC's consumption surfaces.

Layer 1: SAP Analytics Cloud — Consumption and Decision

Optimized Story Experience

Unified responsive story editing and viewing on one canvas, and the only story experience since SAP deprecated the Classic Experience at the end of Q2 2026 (classic stories and analytic applications are no longer accessible; SAP's Story Conversion Tool handles bulk migration). Better mobile and desktop parity.

Seamless Planning with Datasphere

The fact and master data of an SAC planning model are stored in a Datasphere space (chosen per model), while SAC keeps the model structure, calculations, and planning processes. No separate import step: plan data lives next to governed actuals. It needs SAC on SAP HANA Cloud linked 1:1 to a Datasphere tenant and models in the Optimized Design Experience.

Joule AI Assistant

Joule's analytical insights, powered by SAC's just ask (generally available since Q1 2024), answer natural-language questions on indexed SAC models and Datasphere analytic models; standard-release SAC customers have had them since the Q2 2025 release. AI story generation in SAC was announced at SAP Sapphire 2026 as a private preview. Lowers the barrier for self-service analysis without training.

Smart Predict

Automated classification, regression, and time-series forecasting built into planning and analytics workflows, including predictive planning into plan versions. Reduces need for separate predictive tools.

Design principle: SAC is optimized for exploration, collaboration, and action. It does not manage semantic definition or fine-grained governance. Those concerns belong in Datasphere.

Layer 2: SAP Datasphere — Semantic Data Fabric

Spaces and Organization

Logical containers for modeling work, governed by access controls. Separate business domains while sharing a single semantic layer underneath.

Analytic Models

Fact and dimension models with business logic, attributes, hierarchies, and measures. Acts as the single source of business meaning.

Data Products

Named, owned assets published for reuse. In BDC, SAP-managed data products arrive ready to install, and custom data products are published through the Data Sharing Cockpit from tables in a Datasphere file space. They enable governed reuse across SAC, BW, and external consumers. (Semantic modeling itself happens in views and analytic models in the Data Builder; SAP deprecated the separate Business Builder.)

Data Access Controls and Lineage

Row-level security through data access controls, which filter every reader, including SAC (there are no column-level controls, so restrict columns through views). Lineage and impact analysis from source through modeling to consumption, and space audit logs for reads and changes. Auditable, transparent governance.

Key shift: Datasphere moves semantic definition from dashboards and reports into a dedicated shared layer. This enables consistency and reduces fragmentation when multiple teams build on the same business context.

Layer 3: Business Data Cloud — Platform Unification

The Commercial Bundle

Business Data Cloud packages SAP's data and analytics portfolio into one SaaS offering and adds new capabilities on top (data products, intelligent content, BDC Connect). It bundles:

SAC + Datasphere: The core analytics and semantic layer
SAP HANA Cloud + SAP Master Data Governance: Both generally available as core components of Business Data Cloud as of SAP Sapphire 2026 — the data foundation and the master-data layer (master data further reinforced by the Reltio acquisition, completed May 7, 2026, and SAP Reltio in SAP BDC, generally available since October 2026)
SAP BW, private cloud edition (optional): For organizations modernizing existing BW estates; its Data Product Generator turns BW data into data products
SAP Databricks and SAP Snowflake: Partner engines for data engineering, data science, and AI on SAP data products (SAP Snowflake generally available since May 2026)
Intelligent content: Prebuilt line-of-business packages (for example Finance, Supply Chain, People, Spend, and Revenue Intelligence) on certified SAP-managed data products

All bundled under capacity-unit pricing rather than per-user or per-analytics-item pricing. This changes the commercial conversation from "How many stories will we build?" to "What total semantic and decision capacity do we need?"

BDC Connect: Zero-Copy Integration to External Clouds

Share data products where they live, without copying

BDC Connect shares SAP-managed and custom data products zero-copy with partner platforms, and partners can share data products back into BDC. It needs a BDC subscription and a BDC Connect instance in the same region and hyperscaler as the data product, and SAP meters cross-region data transfer and API requests. Live queries from Datasphere to external platforms are a different pattern: federation through remote tables.

Current and planned connectors (Amazon Athena is planned for H1 2027):

Databricks: GA October 2025
Snowflake: GA May 4, 2026
Google BigQuery: GA July 2026
Microsoft Fabric (Copilot & Power BI): public preview; GA planned for the second half of 2026

Use case: Organizations with data already in cloud data warehouses can share it back into BDC as data products and layer SAP analytics and planning on top without re-ingesting.

Data Products and Data Product Studio

A data product is a reusable asset that combines:

Data (for custom products, tables in a Datasphere file space)
Metadata and documentation
Terms and conditions and access agreements (data product governance, September 2026)
Governance and access policies
Publication lifecycle (listed or delisted)

Data Product Studio, announced at SAP TechEd 2025 with general availability planned for H1 2026, is meant to give business and data teams a self-service interface to discover, publish, and govern data products; as of November 2026, SAP had not confirmed its general availability, and custom data products are published through the Data Sharing Cockpit. Either way, this moves the discipline of reusable semantic assets from "best practice" to "built-in workflow."

Why this matters: Without explicit data product discipline, semantic reuse remains informal. Teams end up with duplicate models, inconsistent definitions, and fragmented trust. BDC shifts this toward governed assets with clear ownership.

Integration and BTP Runtime

Integration Suite: Connecting SAP and Non-SAP Systems

SAP Integration Suite is an SAP BTP service in its own right; its main capabilities for analytics are:

Cloud Integration (formerly CPI)

Pre-built connectors and custom flows for moving data between SAP and non-SAP systems. Event-driven or scheduled pipelines; integration flows can write data into Datasphere through an Open SQL schema.

API Management

Publish and govern APIs, such as Datasphere's OData consumption APIs, through the Developer Hub (formerly API business hub enterprise), including governed MCP servers for AI agents. Enable external applications and workflows to call analytics.

Event Mesh

Asynchronous, event-driven messaging: the Event Mesh capability for low-to-moderate volumes, and advanced event mesh for high volumes. Datasphere and SAC do not subscribe to event topics themselves; a subscriber, such as an integration flow, writes events into Datasphere.

BTP as the Platform Runtime

SAP's analytics and integration services run on SAP BTP (Business Technology Platform), now a core part of the SAP Business AI Platform. Key runtime components:

AI Core: Managed runtime for custom models, with the generative AI hub for access to foundation models. SAC's own AI features ship with SAC rather than running on your AI Core.
Cloud Foundry and Kyma: Application runtimes (Kyma is SAP-managed Kubernetes) for custom microservices and extensions.
SAP Build Process Automation: Low-code workflows, decisions, forms, and RPA for approval workflows and process automation.

Architect's note: Understanding BTP is not optional. The analytics platform is not a standalone appliance—it is a cloud application running on a modular BTP foundation. Extensions, integrations, and custom applications will often reach into BTP services.

How the Layers Talk to Each Other

Data Flow Through the Stack

Inbound: Source systems → replication flows, data flows, or Integration Suite (Cloud Integration) → Datasphere tables → Datasphere modeling (views, analytic models) → data products

Outbound: Datasphere models queried by SAC stories → Joule for discovery → Smart Predict for forecasting → planning stores plan data in Datasphere (seamless planning)

External: BDC Connect shares data products zero-copy with Databricks, Snowflake, and BigQuery (Fabric in preview), and partners share data products back; live queries from Datasphere to external platforms use remote tables (federation) instead

Real-time: Event Mesh carries events between applications; event-triggered actions run in integration flows or SAP Build Process Automation, while SAC refreshes on schedules or through live queries rather than on events

Design Principles That Matter

Semantic consistency is upstream. Datasphere models define business meaning once. SAC, BW, and external consumers all use that same definition.
Governance travels with context. Access controls, lineage, release states, and object versioning are managed in Datasphere. They apply everywhere the semantic layer is used.
BDC Connect shifts the economics. Partner platforms use SAP data products without extraction pipelines (cross-region transfer and API requests are metered), and remote tables let Datasphere query external data without ingest. Datasphere acts as a semantic integration layer, not only as a data warehouse.
Planning is a first-class consumer. With seamless planning, a planning model's fact and master data are stored in a Datasphere space, not in a separate imported copy. This closes the loop: reporting and planning share the same governed foundation.
Data products enable reuse discipline. Without explicit ownership and publication, semantic assets become fragmented and duplicate.

What This Means for Architecture Decisions

For Teams Starting with SAC

You don't need Datasphere on day one. But you should architect SAC content assuming Datasphere will become the semantic layer later. Keep modeling logic minimal in SAC stories. Prepare to move query logic downstream when that capability exists.

For Teams Building Datasphere

Treat Datasphere as the single source of business meaning. All SAC stories should query Datasphere models, not source tables directly. This is not a nice-to-have. It is the design principle that makes governance, consistency, and reuse possible.

For Integration and Data Platform Leads

BDC Connect and Integration Suite are not optional add-ons. Plan for them. If your data lives in Databricks, Snowflake, or BigQuery, BDC Connect lets those platforms share it back into BDC as data products instead of copying it into Datasphere. If you have non-SAP systems, Integration Suite becomes your integration backbone. Both change the economics and architecture conversation.

For Governance and Data Product Teams

Data products are not just technical. They require operational discipline: clear ownership, versioning, SLA definition, and publication workflow. Use the Data Sharing Cockpit and the data product governance features released in September 2026 (terms and conditions, steward approval, access agreements) today, and track Data Product Studio, whose general availability SAP had not confirmed as of November 2026, as part of your semantic governance roadmap.

For Planning and Finance Teams

Seamless Planning with Datasphere means planning models finally sit on the same semantic foundation as reporting. This enables consistent dimensions, hierarchies, and measure definitions across planning and analytics. Invest in this connection early.

Why This Matters Now

Enterprise analytics has shifted from dashboards-as-endpoint to analytics-as-platform. The old model was: ETL → BI tool → dashboard. The new model is: Source systems → Integration (Cloud Integration/Event Mesh) → Semantic layer (Datasphere) → Multiple consumption surfaces (SAC, planning, external APIs, Joule AI, workflow automation). Data products add governance discipline to this flow.

SAP's architecture reflects this shift:

Datasphere as the center: Not "our analytics database" but "the semantic foundation for all analytics work"
SAC as the primary consumption interface: Stories, planning, AI-assisted exploration, not as a replacement for all BI tools but as the native SAP platform for analytics and decision support
BDC as the commercial unit: Not "buy SAC or Datasphere" but "buy the platform bundle and add what you need for your data foundation"
Integration Suite as table-stakes: Real enterprises have data everywhere. Zero-copy sharing with external clouds plus event-driven integration become non-negotiable
BTP as the foundation: All of this runs on a modular cloud platform. Extensions and integrations will use BTP services. Plan for it.

The question is no longer "Should we use SAC or Datasphere?" The question is "What role does each layer play in our analytics and decision-support operating model? And how do we govern data products and semantic reuse across all of them?"

Next Steps for Your Architecture

Immediate (0–3 months)

Map your current analytics and planning landscapes to these three layers. Where does semantic definition live today? Where is it fragmented?
Assess data residency. Are key datasets in Databricks, Snowflake, or BigQuery? If so, BDC Connect changes your architecture.
Audit current SAC content for modeling logic that should move downstream to Datasphere. Plan a migration if you have hundreds of stories.

Medium term (3–6 months)

Prototype Datasphere semantic models for your highest-value business domains. Test seamless planning integration.
Define data product standards: ownership, documentation, versioning, publication workflow. Track Data Product Studio, whose general availability SAP had not confirmed as of November 2026.
Plan Integration Suite architecture for source system connectors, event-driven data flows, and API publication.

Strategic (6+ months)

Migrate high-risk semantic definition from SAC and BW into Datasphere. Establish it as the system of record for business meaning.
Build out the data product catalog and the Data Sharing Cockpit publication workflow. Make semantic reuse a governed, visible process.
Evaluate Joule AI and Smart Predict in your own data context. These are not late-stage features but core consumption patterns.
Plan for BTP extensions and custom microservices. The platform is modular. What custom analytics, workflow, or integration logic belongs in BTP vs. above it?

When this article applies

Use this when you need to explain to stakeholders, architects, or new team members how SAC, Datasphere, and Business Data Cloud relate as a unified platform. This article covers architecture and strategy but not detailed product configurations.

When you need to go deeper

If you need help with specific product capabilities, comparison matrices, migration strategies, or detailed BTP design, move to the dedicated topic pages for SAP Analytics Cloud, SAP Datasphere, integration patterns, or BTP architecture.

What to read next

Start with "What is the SAP analytics landscape in 2026?" for context. Then dive into "SAP Analytics Cloud architecture" or "Datasphere semantic layer design" depending on which layer you are designing. Finish with "BTP and SAP analytics" to understand the runtime.

References and Further Reading

Attribution and sources: This article synthesizes information from SAP's official learning courses, product documentation, and public roadmaps as of November 2026. Specific release dates and capability scopes should be verified against official SAP communications before using for high-stakes decisions. This content is independent editorial analysis, not official SAP product documentation.