OneLake is the storage layer under Microsoft Fabric: one copy of your data that every tool reads without duplicating it. What that means in plain English.
Microsoft OneLake is the single storage layer underneath every Microsoft Fabric workload: one lake for the whole organisation, where each table is stored once in the open Delta Parquet format and read directly by pipelines, notebooks, the warehouse, Real-Time Intelligence and Power BI without being copied between them. It works like OneDrive for data: every workspace gets a place in the same lake, shortcuts point at data that lives elsewhere (another workspace, Azure Data Lake Storage, Amazon S3, Google Cloud Storage or Databricks) without moving it, and mirroring keeps operational databases such as Azure SQL, Snowflake and, once the preview arrives, Business Central replicated into it. This post explains the idea in one sentence, why one copy matters, and what a shortcut actually is.
The idea in one sentence
OneLake is the single storage layer that sits underneath Microsoft Fabric. It is one place where all your data lives, that every Fabric tool can read without copying it first. Microsoft describes it as the OneDrive for data, and the comparison is fair. You have one store, and everything else plugs into it.
Why one copy matters
For years the default pattern was to copy. The warehouse had its copy, the lake had another, each reporting team pulled its own extract, and before long the same number existed in five places that quietly disagreed. Every copy is a cost, a delay and a chance to be wrong. OneLake is built around the opposite idea. The data sits once, and the tools come to it. A report, a notebook, a pipeline and a model can all point at the same table without anyone duplicating it.
What it actually is
Under the surface OneLake is one logical data lake for your whole tenant, built on the open Delta and Parquet formats. Open matters here. Because the data is stored in a format other tools understand, you are not locked into reading it only through Microsoft. The storage is shared, the format is standard, and access is governed centrally rather than per copy.
Shortcuts, so you do not have to move data
One of the most useful ideas in OneLake is the shortcut. A shortcut is a pointer to data that lives somewhere else, another lake, another cloud, another workspace, that appears in OneLake as if it were local. You query it in place. That means you can bring data together for analysis without a migration project and without making yet another copy that someone has to keep in sync. For a lot of real problems, a shortcut is the whole answer.
What this changes for you
The practical effect is less plumbing and more trust. Fewer copies means fewer pipelines to maintain, fewer windows where numbers drift apart, and a clearer answer to the question every data team dreads, which is which version is right. It also makes governance simpler, because you are securing one store rather than chasing permissions across a dozen extracts. None of this is magic, and a OneLake left untended can still sprawl. But the default it encourages, one copy and many tools, is the right one.
Where to start
You do not adopt OneLake as a project in its own right. It comes with Fabric, and the useful move is to stop copying by habit. Next time someone is about to extract a dataset to work on it, ask whether a shortcut would do instead. That one question, asked often, is how an estate gets simpler rather than messier over time. If you want help working out where OneLake fits in what you already run, that is a good conversation to have.
Related reading
Fabric capacity: a plain-English guide to not getting a shock on your invoice
Kiran Sahiba
Analytics Solutions Consultant
Part of the Hopton Analytics team, delivering governed analytics programmes for UK mid-market organisations.
