# Microsoft Fabric Warehouse

> SaaS T-SQL warehouse storing Delta tables in OneLake

Microsoft Fabric Warehouse is the relational warehouse item in Microsoft Fabric, offering full T-SQL with multi-table ACID transactions over Delta tables held in OneLake and billed against a shared Fabric capacity.

Website: https://www.microsoft.com/en-us/microsoft-fabric

## Use it when

- The organization is standardized on Power BI and Microsoft; the warehouse, lakehouse, and semantic models share one SaaS platform and one copy of the data in OneLake.
- You want full T-SQL with multi-table ACID transactions over open Delta tables rather than a proprietary store.
- Azure Synapse dedicated SQL pools are your present; this is where Microsoft points that migration.
- Zero-knob operation is the preference: workload management, statistics, and compaction are automatic.

## Think twice when

- You are not otherwise a Microsoft shop; the value concentrates when Entra ID, Power BI, and Fabric capacities are already there, and it is Azure only.
- Streaming ingest into the warehouse item is not native; Eventstream lands in a lakehouse, and the warehouse loads by batch.
- Capacity-based billing pools all Fabric workloads against one F SKU; a noisy neighbor conversation replaces a per-warehouse bill.

## How it runs

SaaS within Microsoft Fabric: serverless compute nodes are assigned from your capacity, storage is Delta tables in OneLake, and both scale independently with nothing to size. Spark notebooks and Power BI read the same files. Loading goes through COPY INTO, pipelines, dataflows, or CTAS; billing smooths Capacity Unit-seconds against the SKU.

## Details

### General

| Attribute | Value |
| --- | --- |
| License | Commercial |
| Deployment | Managed, Serverless |
| Operational complexity | Low |
| Pricing | Subscription |
| Documentation | https://learn.microsoft.com/en-us/fabric/data-warehouse/ |

### Data Warehouses

| Attribute | Value |
| --- | --- |
| Where it runs | Azure only |
| Compute and storage | MPP over OneLake: serverless nodes assigned from capacity, nothing to size |
| Storage format | Open: Delta tables in OneLake |
| SQL dialect | T-SQL |
| Streaming ingest | Batch only: COPY INTO, pipelines, CTAS; Eventstream lands in a lakehouse |
| Concurrent users | Autoscaled by capacity SKU: burst up to 12x baseline, SELECT and ingest pools isolated 50/50 |
| Acceleration | V-Order writes, automatic statistics and compaction, transparent memory and SSD cache; clustering keys at create time |
| Billing unit | Capacity Unit-second against an F SKU, smoothed; OneLake storage billed separately |
| Time travel | 1–120 days, 30 by default |
| Access control | Entra ID with workspace roles and item permissions; object, column and row-level security, dynamic data masking |

## Capabilities

- [Data Warehouses](https://matca.io/capabilities/data-warehouse)

---

Source: https://matca.io/microsoft-fabric-warehouse
Last updated: 2026-08-31T11:40:47.586Z