Microsoft Fabric pricing explained: what it costs and how to control it

The pricing model in one paragraph

Microsoft Fabric is sold by capacity, not by user. You buy a unit of compute called a Fabric capacity (measured in capacity units, or CUs), and everyone in your organisation can use it. The smallest capacity (F2) starts at roughly £200 per month on pay-as-you-go, and you can scale up or down as your needs change. Storage in OneLake is billed separately, at a low per-gigabyte rate. There are no per-user Fabric licences for most workloads, although users who need to author Power BI content still need a Power BI Pro licence.

Pay as you go vs reserved capacity

You can buy Fabric capacity on a pay-as-you-go basis, billed by the second, or commit to a reserved capacity for a year and save around 40 percent. Pay as you go suits businesses that are still working out their usage patterns or want to pause capacity outside working hours. Reserved capacity is the right choice once your workload is predictable. The good news is you can start on pay as you go, learn what you actually use, then move to a reserved commitment without changing anything else.

Realistic monthly costs

A small team running reporting and modest data engineering tends to be comfortable on an F2 or F4 capacity (£200 to £400 a month). A mid-market business with several departments using Fabric for analytics, data warehousing, and Copilot typically sits in the F8 to F16 range (£800 to £1,600 a month). Larger enterprises with real-time analytics, data science workloads, and substantial pipelines move into F32 and above. These figures exclude storage and any Power BI Pro licences, which are usually a small fraction of the total.

Where bills tend to creep up

The most common cause of unexpected costs is leaving capacity running 24/7 when your workload only needs it during business hours. Pausing the capacity outside core hours can cut costs by more than half. The second common cause is uncontrolled data pipelines that refresh more often than they need to. A dashboard refreshing every 15 minutes when daily would do is a quiet but expensive habit. The third is overprovisioning. It is tempting to buy a larger capacity for safety, but Fabric scales fast, so it is usually cheaper to start small and scale up when you see real demand.

How to keep it predictable

Three practices keep Fabric bills sensible. First, monitor capacity usage from day one using the built-in Capacity Metrics app. Second, set up alerts so you know before you hit your ceiling. Third, review your pipelines and refresh schedules quarterly. None of these are technical heavy lifting, but they are easy to skip when nobody owns them.

The hidden costs people forget to budget for

The capacity SKU is the line item everyone sees, but two costs catch businesses out. The first is Power BI Pro licences for anyone authoring content - even on F64 and above, where viewers do not need their own licence, authors still do, and at around £8 per user per month that can add up if you have a wide analyst pool. The second is data egress: if your source data lives outside Microsoft's cloud (in AWS, on-prem, or in another SaaS platform), moving it into OneLake can generate transfer charges. Both are usually small compared to the capacity itself, but they are worth modelling rather than discovering on the invoice.

Getting the most out of your spend

The biggest cost saving with Fabric is not in the licensing, it is in the consolidation. Most businesses adopting Fabric retire two or three other tools at the same time, which often pays for the platform outright. To make that work, the data going into Fabric needs to be in good shape, which is where our data readiness work comes in. The Microsoft Fabric page has more on how we help businesses get deployed without overspending.

Want a straight answer on what Fabric would cost your business? Talk to us.