Planning in Microsoft Fabric IQ is about to enter general availability. Microsoft introduced it as a public preview at FabCon Atlanta in March and has said GA lands later in June, which is fast for an enterprise planning and write-back workload. (the Fabric blog announcement)
One correction I keep making, since people place it at Build: the reveal was FabCon. Build 2026 in early June is where the confusion starts, because Microsoft positioned Fabric IQ broadly as GA there and put Planning close behind, so the two get merged. (Microsoft’s phrasing for Planning is “later in June,” and no firm date is on the page as I write this.)
I have already written what Planning is and how to stand up your first item. Those links are at the bottom. The questions I get now are the practical two. What does it cost, and can I run it where my data already lives.
There is no per-seat license. Planning runs on your Fabric capacity and consumes capacity units, the same pool that already runs your Power BI and the rest of Fabric. You do not buy anything on the side.
What trips people up is how the meter works. The first action a user takes sets their role, and that role’s capacity is billed for the next thirty days. Build or model the plan and you are a Planner. Enter or write back data and you are a Stakeholder. View only and you are a Viewer.
Each role holds a steady slice of capacity, measured in the same unit your Fabric capacity (the F-SKU) is sold in. A Planner holds 1.16 CU, a Stakeholder 0.23 CU, a Viewer 0.05 CU. That is the share of your capacity the role keeps occupied, so it lines up directly with the SKU number: one Planner is 1.16 of an F2’s 2 CU, about 58 percent of it. The roles sit about five times apart at each step, so a Viewer is close to free and most of your cost is the small group actually building.
Add the slices across everyone and that total is the capacity your plan needs. A worked example from the calculator I built: three Planners, twenty Stakeholders, and fifty Viewers come to 10.58 CU, or 12.17 with a 15 percent buffer on top. That fits an F16, about three quarters full.
Figure 1. The capacity calculator: three Planners, twenty Stakeholders, and fifty Viewers, with a 15 percent buffer, size to an F16 at about three quarters full.
Turn on dollars and pick a region. The CU each role consumes is the same everywhere; the region only changes the price of one CU. In Central US that F16 runs about $2,100 a month pay-as-you-go, or about $1,250 on a one-year reservation, roughly 41 percent less. About $1,600 of the pay-as-you-go bill is Planning’s share, the same three quarters your plan occupies on the F16, and the rest is headroom for other Fabric work. (the calculator splits the one capacity bill into those parts; they are not separate charges.)
Figure 2. The same plan in dollars (Central US): the F16 pay-as-you-go next to the one-year reservation, with the per-role breakdown.
The calculator keeps to one unit and one set of numbers, current as of June 2026. You can run your own user mix in it at fabricplanning.io/calculator. For the live dollar cost of a CU in your region, check Microsoft’s Fabric pricing page, and for a fuller estimate, Lumel, Microsoft’s development partner for Planning.
Can you run it in your region
Before you create your first Plan item, check one thing: your region.
In the preview region list, East US and East US 2 are both excluded, along with about a dozen others. The likely reason is the current scramble for AI compute: those regions are heavily backlogged, and there is no published date for when Planning reaches them. If you are in the East, treat the timeline as unknown.
If you are blocked and cannot wait, there is a workaround, though it costs you. Spin up a separate Fabric capacity in a supported region (North Central US, for example) and run the Plan there, while you keep the Fabric SQL database that holds your plan data in your source region. OneLake shortcuts bridge the two without copying the data. You manage that Plan capacity, and its cost, on its own. You pay for the path in cross-region latency and some data egress. And if residency rules keep your data in an East datacenter, get security to sign off first, because the write-back still crosses the line.
Treat the cross-region setup as a bridge. You cannot change a workspace’s region in place, so when your home region reaches GA you migrate the Plan over. The Plan item supports source control, so that move is a redeploy into the new region rather than a rebuild. And verify availability with Microsoft before you pick, because choosing a region in any tool does not confirm that Planning runs there.
Pick the region you can run in today, and schedule the move home for when Planning reaches yours.
Want to go from reading to building?
PowerTable is where I tell everyone to start with Fabric Planning. It is the lowest-risk way in, and it is the part I use most. If you want to go from reading about it to building in it, I am running a hands-on workshop on exactly that.
Mastering Fabric Planning PowerTable is a two half-day virtual workshop (June 29 and 30, 2026), taught with four other Fabric and FP&A practitioners: Nikola Ilic, Shabnam Watson, Artur König, and Miguel Félix. It is a small working cohort with direct Q&A, not a webinar (it is hands-on and assumes you are comfortable with Fabric basics). The first cohort is half off, $499 instead of $999.
Save your seat for Mastering Fabric Planning PowerTable.
Fabric Planning is my field guide to planning and budgeting going native in Microsoft Fabric, written from the FP&A side of the table. Subscribe to get the next one.
Further reading: “Fabric Planning: What, So What, and Now What?” and “Fabric Plan Step Zero.”
An independent knowledge hub for Microsoft Fabric Planning. Not affiliated with Microsoft. Published by Data Crafters, a Microsoft Fabric Featured Partner.