Every week we get the same question from a client: "Microsoft keep pushing Fabric — do we actually need it for Power BI?" The honest answer for most UK mid-sized businesses is no, not yet, and possibly not ever. Power BI on its own continues to work exactly as it always has. Fabric is a useful addition for specific situations, not a prerequisite. Here is how to tell which side of the line you sit on.
When Power BI on its own is the right answer
Power BI Pro at roughly £8 per user per month, or Premium Per User at £18, covers the vast majority of UK business reporting needs. You probably do not need Fabric if all of the following are true:
Your largest semantic model fits comfortably under a few gigabytes and refreshes in under an hour. Your data sources are well-behaved databases, SharePoint lists, or a manageable handful of CSVs. You do not need to share storage between Power BI and other tools like notebooks or warehouses. You have fewer than two or three hundred report consumers. And nobody is asking for near-real-time data.
If that describes your setup, Fabric will add cost and complexity without giving you anything you cannot already do.
The four situations where Fabric earns its keep
1. You have outgrown the Power BI import-model size limit. Pro caps datasets at 1 GB; PPU at 100 GB. When you start running into refresh timeouts, gateway memory errors, or dataset eviction, Fabric's DirectLake mode lets Power BI query Delta tables in OneLake without importing them at all. Genuinely transformative for big data.
2. You are stitching several Azure services together by hand. If you already have Data Factory, Synapse, Databricks and Power BI all running separately and somebody spends their week wiring them together, Fabric collapses most of that into one platform with one bill. The simplification alone is often worth it.
3. You need a lakehouse without a data platform team. Mid-sized businesses that want a proper modern data architecture but cannot justify hiring three data engineers benefit from Fabric's everything-in-one-place model. The total cost of ownership is far lower than building the same thing on raw Azure.
4. You have several Power BI workspaces pulling from the same data. When five different teams all import a copy of the sales table into their own dataset, you end up with five subtly different versions of the truth. A Fabric Lakehouse with shortcuts lets everyone query the same physical data without duplicating it.
The hidden cost of jumping in too early
The cheapest Fabric capacity (F2) costs roughly £210 per month in the UK if you leave it running. That is the equivalent of about 12 PPU licences. So if you have fewer than a dozen heavy report users, switching to Fabric to "future-proof" usually costs more than it saves.
There is also a hidden cost in expertise. Fabric introduces concepts (capacity units, smoothing, OneLake shortcuts, DirectLake fallback modes) that someone on your team or your consultancy needs to understand to avoid runaway bills. If nobody on the project has shipped a Fabric workload before, the first three months will be more expensive than the licence.
A simple decision framework
Answer these three questions honestly:
Are your Power BI refreshes regularly failing or taking longer than the business can tolerate? If no, you almost certainly do not need Fabric. If yes, Fabric or Premium might help, but so might fixing the underlying model.
Do you have data that needs to be shared between Power BI and at least one other workload (a notebook, a warehouse, a real-time feed)? If no, stay on Power BI. If yes, OneLake is genuinely the cleanest solution on the market.
Do you have someone — internal or external — who knows Fabric well enough to keep capacity costs under control? If no, do not adopt Fabric until you do.
Migration is not a one-way door
The reassuring thing about Fabric is that adopting it does not force you to throw anything away. Existing Power BI workspaces can live alongside Fabric workspaces in the same tenant. You can pilot a single workload on F2 capacity for a month, then either scale up or quietly retire it. We routinely run small Fabric proofs of concept that cost a few hundred pounds end to end.
If you want a steer on whether your specific situation is one Fabric would help, our Microsoft Fabric consultancy page describes how we usually run a short evaluation, and our plain-English Fabric guide covers the basics in more detail.
Frequently asked questions
Can I use Power BI without Microsoft Fabric?
Yes. Power BI Pro and Premium continue to work exactly as before. Fabric is an optional addition, not a replacement.
Will Microsoft stop selling Power BI on its own?
There is no announcement that suggests so. Power BI Pro and PPU are still actively sold and supported. Fabric simply makes Power BI a workload within a wider platform for customers who want that.
When is the right time to move to Fabric?
When you have a concrete pain Fabric solves — large datasets, multiple Azure services to consolidate, a need to share storage across workloads — and someone who can manage capacity costs. Not before.
Want to talk this through with someone?
We are an independent UK Power BI and Microsoft Fabric consultancy. Honest opinions, fair prices, no sales pressure.