How to Get Started with Microsoft Fabric

Microsoft Fabric is a unified analytics platform that consolidates data ingestion, storage, transformation, and reporting under a single Microsoft-hosted workspace. For teams already using Power BI or Microsoft 365, it replaces the need to manage Azure Data Factory, Azure Synapse Analytics, and Power BI as separate services. A 60-day free trial is available at app.fabric.microsoft.com. This guide covers how to activate a workspace, create a lakehouse, load data, and publish your first Power BI report.

What Microsoft Fabric Is

Microsoft launched Fabric in May 2023 and made it generally available in November 2023. By early 2026, more than 70 percent of Fortune 500 companies had at least one Fabric workload running in production, according to Microsoft FY2025 earnings disclosures. The platform runs on OneLake, a single logical storage layer shared across all services in a workspace.

OneLake stores every table and file in open Delta Parquet format. Data written by a Fabric pipeline is immediately readable by Spark, DuckDB, Polars, or Databricks without any export step. The practical effect is that you can build a data warehouse inside Fabric while still querying it from an external Python environment or notebook tool.

Fabric bundles six product experiences under a single capacity license: Data Engineering for Spark notebooks and batch jobs, Data Factory for low-code data ingestion pipelines, Data Science for machine learning notebooks and model tracking, Data Warehouse for T-SQL warehousing over OneLake, Real-Time Intelligence for event streams and KQL databases, and Power BI for reporting. For most analysts and operations teams starting out, the relevant three are Data Factory, Lakehouse, and Power BI.

Starting Your Free Trial

Go to app.fabric.microsoft.com and sign in with a Microsoft 365 account. If your organization already holds a Power BI Pro or Premium Per User license, Fabric may already be available in your account. Otherwise, click the account icon in the top-right corner and select Start trial. The trial lasts 60 days and includes access to all Fabric workloads at no charge.

No Azure subscription is required. Fabric billing runs through Microsoft capacity units rather than Azure billing lines, which matters for organizations operating outside Azure. If your IT administrator has disabled self-service trials, sign-in returns an access error. The admin can re-enable trials at admin.microsoft.com under Purchase Services.

Creating a Workspace

Workspaces are Fabric's top-level containers. They hold lakehouses, warehouses, pipelines, notebooks, and reports together, and they determine which capacity unit is assigned to run your work.

In the Fabric home screen, click Workspaces in the left navigation sidebar, then select New Workspace. Name it (for example, Sales Analytics), set the License mode to Trial, and click Apply. The workspace appears in the sidebar and is backed by shared trial capacity. For production workloads you would assign dedicated F-SKU capacity, but trial capacity is sufficient for evaluation.

Creating a Lakehouse

A lakehouse is the primary storage object in Fabric. It has two sections: Tables, which hold Delta-format data queryable via SQL or Spark, and Files, which hold raw files in any format before transformation.

Inside your workspace, click New item, select Lakehouse, name it, and click Create. Fabric provisions the lakehouse and automatically creates a SQL analytics endpoint that exposes the Tables section for read-only SQL queries.

Loading Data

Fabric provides three practical ingestion paths for teams starting out.

The most direct is a file upload. In your lakehouse, click Upload in the Files section and drag in a CSV or Parquet file. The file is stored raw. To make it a queryable Delta table, right-click the file and select Load to Tables. Fabric converts the CSV to Delta format without any additional configuration.

For recurring ingestion from a SaaS application, use a Data Factory pipeline. Click New item, select Data pipeline, add a data source connector (Fabric ships with more than 150, including Salesforce, HubSpot, SharePoint, and Google Sheets), and set the destination to your lakehouse Tables section. Pipelines run on a schedule you configure, typically hourly or daily.

If your data already lives in Azure Data Lake Gen2, Amazon S3, or Google Cloud Storage, create a Shortcut instead of copying the data. From the lakehouse Files section, click New shortcut and point to the external storage path. Fabric presents the external data as if it were local without transferring it, avoiding duplicate storage costs.

Querying Data with SQL

Once data appears in the Tables section, switch to the SQL analytics endpoint view using the selector in the top-right corner of the lakehouse. A standard SQL editor connects directly to your Delta tables.

A basic query looks like this:

SELECT TOP 100 * FROM SalesData.dbo.your_table_name;

The SQL endpoint is read-only. Writes to Delta tables go through Spark notebooks or pipelines. This surprises analysts accustomed to warehouses where SQL handles both reads and writes. If you need full SQL write access, create a Fabric Data Warehouse item instead of a lakehouse. Warehouses support INSERT, UPDATE, and DELETE via T-SQL and are better suited for teams with existing SQL-heavy transformation workflows.

Building a Power BI Report

Fabric's most useful integration for business users is the direct path from lakehouse to Power BI report in the browser, without installing Power BI Desktop or configuring a data gateway.

In the SQL analytics endpoint view, click New report in the top toolbar. Power BI opens in the browser with a live semantic model connection to your lakehouse tables. Drag fields onto the canvas to add visuals, apply filters, and format charts, then click Save to publish the report to your workspace.

Data updates in the lakehouse flow into the report on the next scheduled refresh. No separately published datasets, on-premises data gateways, or .pbix files are required, which significantly reduces setup time compared to traditional Power BI workflows that start from a local desktop connection.

Understanding Capacity Pricing

Fabric bills on compute capacity units rather than per query. As of 2026, an F2 dedicated unit costs approximately $262 per month and covers all workloads running in assigned workspaces. If running workloads exceed available capacity, jobs queue rather than fail outright.

This pooled model differs from per-query pricing in Snowflake or BigQuery, where costs scale directly with query volume. For teams with predictable scheduled pipelines and daily report refreshes, Fabric's flat capacity pricing is more cost-predictable. For teams with sporadic heavy query bursts, per-query pricing can be more economical.

Common Pitfalls

Spark defaults to memory-heavy cluster configurations even for small data. Running a notebook that reads a 10 MB file allocates the same starter Spark cluster as one processing 10 GB. Use the SQL analytics endpoint for lightweight queries to avoid unnecessary capacity consumption.

Delta Parquet is an open standard, but legacy reporting tools built on ODBC connections may not support it directly. Check connector compatibility before migrating an existing reporting stack from a traditional SQL Server or Oracle source.

Trial capacity is shared across Microsoft's multi-tenant infrastructure and may show slower execution than dedicated capacity, particularly during peak hours. Do not benchmark trial performance against production SKU expectations.

When Fabric Makes Sense

Fabric works best for organizations already using Microsoft 365, where Power BI is the reporting layer and Excel is the primary analyst interface. It consolidates multiple Azure billing lines into one capacity fee and reduces the number of connection credentials to manage across services.

For teams outside the Microsoft ecosystem, or those requiring multi-cloud portability and open-source orchestration, Databricks combined with dbt offers more flexibility. For teams that need immediate analysis from files they already have without provisioning any infrastructure, tools like VSLZ let you upload a dataset and get summaries, statistical analysis, and charts from a single prompt with no warehouse setup required.

Summary

Getting started with Microsoft Fabric takes around 30 minutes from sign-up to first report. Activate a free trial at app.fabric.microsoft.com, create a workspace, add a lakehouse, load data via file upload or pipeline, query the SQL analytics endpoint, and publish a Power BI report directly in the browser. Plan around capacity unit pricing before moving production workloads from standalone Azure services, and verify connector compatibility for any legacy reporting tools before committing to a full migration.