Microsoft Fabric Pricing and Licensing Explained for Australia
Microsoft Fabric Pricing and Licensing Explained for Australia
Microsoft Fabric pricing is one of the most common questions we get from Australian businesses evaluating the platform. The model is different from what most teams are used to with Azure Synapse or standalone Power BI Premium, and the lack of clear AUD pricing on Microsoft's website doesn't help.
In this post, I'll break down how Fabric licensing actually works, what you can expect to pay in Australian dollars, and where we've seen organisations either overspend or underestimate their requirements.
How Microsoft Fabric Licensing Works
Microsoft Fabric uses a capacity-based licensing model. Instead of paying per user or per service, you purchase a pool of compute capacity measured in Capacity Units (CUs). All Fabric workloads - Data Engineering, Data Warehousing, Real-Time Analytics, Data Science, and Power BI - draw from this single capacity pool.
This is a significant shift from the old model where you might have separate licences for Azure Synapse, Azure Data Factory, Power BI Premium, and Azure Machine Learning. Fabric consolidates these into one billing model.
There are two main ways to purchase Fabric capacity:
Fabric Capacity (Pay-as-you-go) - Billed hourly through your Azure subscription. You can pause and resume capacity, which means you only pay when it's running.
Fabric Capacity (Reserved) - One-year or three-year reservations that offer significant discounts over pay-as-you-go pricing.
Power BI Premium per Capacity (P SKUs) - If you already hold P1 or higher Power BI Premium licences, these include Fabric capacity at no additional cost.
Microsoft Fabric SKU Tiers and AUD Pricing
Here's where it gets practical. Microsoft Fabric capacity comes in SKU tiers, each with a set number of Capacity Units. Pricing in Australia typically runs 10-15% higher than USD list prices due to regional adjustments and GST.
As of early 2026, approximate AUD pricing for Fabric capacity (pay-as-you-go, per hour) looks like this:
| SKU | Capacity Units | Approx. AUD/Hour (Pay-as-you-go) | Approx. AUD/Month (24/7) |
|---|---|---|---|
| F2 | 2 CU | ~$0.55 | ~$400 |
| F4 | 4 CU | ~$1.10 | ~$800 |
| F8 | 8 CU | ~$2.20 | ~$1,600 |
| F16 | 16 CU | ~$4.40 | ~$3,200 |
| F32 | 32 CU | ~$8.80 | ~$6,400 |
| F64 | 64 CU | ~$17.60 | ~$12,800 |
| F128 | 128 CU | ~$35.20 | ~$25,600 |
| F256 | 256 CU | ~$70.40 | ~$51,200 |
| F512 | 512 CU | ~$140.80 | ~$102,400 |
Note: These are approximate figures. Actual pricing varies based on your Microsoft Enterprise Agreement, region, and any negotiated discounts. Always confirm with your Microsoft account team or licensing partner.
Reserved instances (one-year commitment) typically offer 25-35% savings over pay-as-you-go rates. Three-year reservations push that to 40-50%.
What About Power BI Pro and Premium Per User?
Fabric capacity is separate from Power BI Pro and Premium Per User (PPU) licences. Here's how they fit together:
Power BI Pro (~$16.90 AUD/user/month) - Required for sharing reports and dashboards. Every user who consumes or creates Power BI content in a shared workspace needs at least a Pro licence unless you're using Fabric/Premium capacity for distribution.
Power BI Premium Per User (~$33.80 AUD/user/month) - Gives individual users access to Premium features like paginated reports and larger datasets, but does not include Fabric capacity.
Fabric Capacity (F SKUs) - When you deploy content to a Fabric capacity workspace, viewers only need a free Power BI licence. This is where the economics start to make sense for organisations with large numbers of report consumers.
The crossover point we typically see is around 300-500 report viewers. Below that, Power BI Pro licences for everyone may be cheaper. Above that, a Fabric F8 or F16 capacity with free viewer licences often wins on cost.
How to Estimate Your Monthly Fabric Spend
We use a simple framework when helping clients estimate costs:
Step 1: Identify your workloads
List every data workload you plan to run on Fabric. Common ones include:
- Data ingestion (Data Factory pipelines)
- Data transformation (Spark notebooks or Dataflows Gen2)
- Data warehousing (Fabric Warehouse or Lakehouse SQL endpoint)
- Reporting (Power BI semantic models and reports)
- Real-time analytics (KQL databases and eventstreams)
Step 2: Estimate peak and average CU consumption
Each workload type consumes CUs at different rates. Data engineering jobs with Spark tend to be the heaviest consumers. Power BI report rendering is typically lighter but sustained throughout business hours.
Microsoft provides a Capacity Metrics app that shows CU consumption in real time. We always recommend starting with a trial or dev/test capacity to measure actual consumption before committing to a production SKU.
Step 3: Factor in burst and smoothing
Fabric uses a "smoothing" mechanism that spreads CU consumption over time. Short bursts above your capacity limit won't immediately throttle you. But sustained overconsumption will result in throttling or rejection of background jobs.
Step 4: Right-size your SKU
In our experience, most mid-market Australian companies (500-5,000 employees) land somewhere between F8 and F64 for production workloads. That puts monthly spend in the range of $1,600 to $12,800 AUD on pay-as-you-go, or roughly $1,000 to $8,500 AUD with reserved pricing.
Where We See Australian Companies Overspend
A few patterns come up repeatedly in our consulting work:
Running capacity 24/7 when it's not needed. If your Fabric workloads are primarily batch ETL that runs overnight and reporting that's consumed during business hours, you may be able to pause capacity during off-peak hours. This alone can cut costs by 30-50%.
Over-provisioning "just in case." Starting with an F64 when an F16 would handle your workloads is a common mistake driven by uncertainty. We recommend starting smaller and scaling up based on actual Capacity Metrics data.
Ignoring Spark job optimisation. Poorly written Spark notebooks can consume 5-10x more CUs than optimised ones. Investing in code quality and job tuning pays for itself quickly.
Not consolidating Power BI licences. Some organisations keep paying for hundreds of Power BI Pro licences alongside Fabric capacity, when moving those users to free licences on Fabric workspaces would save thousands per month.
Where We See Companies Underestimate Costs
On the flip side:
Real-time analytics is CU-hungry. If you're planning to run KQL databases with continuous event ingestion, the CU consumption can be significant. Budget for at least F32 if real-time is a primary use case.
Data warehouse query concurrency matters. The Fabric Warehouse performs well for moderate query loads, but high-concurrency reporting scenarios (50+ concurrent users running ad-hoc queries) can push you into higher SKU tiers faster than expected.
Development and testing need their own capacity. Running dev/test workloads on your production capacity is risky and can affect performance. Budget for a separate, smaller capacity for non-production use.
Fabric Pricing Compared to the Old Azure Analytics Stack
One of the strongest arguments for Fabric is cost consolidation. In the pre-Fabric world, a typical mid-market analytics stack might look like:
| Service | Typical AUD/Month |
|---|---|
| Azure Synapse Dedicated Pool (DW100c) | ~$1,800 |
| Azure Data Factory (moderate pipelines) | ~$500-1,500 |
| Power BI Premium P1 | ~$7,500 |
| Azure Databricks (small cluster) | ~$2,000-5,000 |
| Azure Storage (data lake) | ~$200-500 |
| Total | ~$12,000-16,300 |
With Fabric, the same workloads might run on an F32 or F64 capacity at $6,400-$12,800/month (pay-as-you-go) or less with reservations. Storage in OneLake is billed separately but is relatively cheap (similar to Azure Data Lake Storage Gen2 pricing).
The savings aren't guaranteed - it depends entirely on your workload profile - but we've seen 20-40% cost reductions for organisations that were previously running multiple Azure analytics services.
Licensing Gotchas for Australian Organisations
A few things to watch out for:
GST applies. All Azure consumption in Australia includes 10% GST. Make sure your budget comparisons are apples-to-apples.
Enterprise Agreement vs. Pay-as-you-go. If you're on a Microsoft Enterprise Agreement, you may get better rates than the public pay-as-you-go pricing. Talk to your Microsoft licensing partner.
Fabric trials are generous. Microsoft offers a 60-day Fabric trial with a reasonable amount of capacity. Use it. We recommend every organisation considering Fabric runs a proof of concept on the trial before committing budget.
Capacity can be shared across workspaces. You don't need separate capacities for each team or project. A single F32 can serve multiple workspaces, which helps with utilisation and cost efficiency.
How Team 400 Helps with Fabric Licensing Decisions
Getting Fabric licensing right is one of the first things we work through with clients. Our Microsoft Fabric consulting engagements typically start with a workload assessment where we measure your current analytics consumption, map it to Fabric CU requirements, and recommend a SKU tier.
We also help with:
- Migration planning from Azure Synapse and Data Factory to Fabric
- Power BI optimisation to reduce CU consumption
- Capacity monitoring and right-sizing after go-live
- Cost modelling for broader AI and analytics initiatives
If you're evaluating Fabric and want a clear picture of what it will cost for your organisation, get in touch with our team. We'll give you a straight answer based on your actual workloads, not a generic estimate.
You can also explore our full range of data and AI services to see how Fabric fits into a broader analytics strategy.