Microsoft Fabric capacity costs: build a budget you can explain line by line.

• Capacity: one or more capacity SKUs billed per hour while active. This is usually the largest line and the one most sensitive to schedule.
• Schedule: capacities can be paused. A capacity active 10 hours on weekdays costs a fraction of one running around the clock, so the active-hours assumption must be explicit.
• Storage: OneLake storage is billed separately from compute and grows with retention decisions.
• Licensing: user licenses for authoring and consumption depend on capacity size and on which workloads people use.
• Networking and adjacent services: data movement, private endpoints, and any Azure resources the platform depends on.

A budget line without its assumption cannot be reviewed. For each line, record the rate used and where it came from, the quantity, and the schedule. When a stakeholder asks why the figure changed between drafts, the assumption that moved is visible. This is the discipline StackSpend Lab enforces: one capacity rate, an active-hours schedule, and other costs, each traceable in the scenario output.

• Rate: the list price per capacity unit-hour for the chosen region, with the date it was read.
• Schedule: active hours per weekday and per weekend day, plus the holidays the capacity stays paused.
• Storage: current volume in OneLake and the expected monthly growth, with retention decisions noted.
• Users: counts per license type, and which workloads those users actually touch.
• Adjacent services: every Azure resource the platform depends on, listed even when the cost is small.

Presenting three or four scenarios side by side moves the conversation from 'is this number right' to 'which operating pattern do we want', which is the decision the business actually owns.

• Baseline: the capacity size the workload needs during business hours, paused overnight and at weekends.
• Always-on: the same capacity without pause, which is what a naive quote often assumes.
• Right-sized: a smaller capacity with longer active hours, or a larger one with shorter windows, to see which the workload tolerates.
• Growth: the baseline with storage and user counts projected twelve months out.

A scenario built from list rates is planning arithmetic. It is not a vendor quote, not a savings promise, and not a substitute for measured consumption once the platform runs. Treat the model as the frame for the first quarter's actuals: compare billed hours and storage to the assumptions monthly and update the model rather than the narrative.

Ascendant Systems helps teams turn a Fabric or broader cloud budget into an operable plan: environment design, pause and resume automation, cost visibility in the same dashboards the business reads, and the recovery readiness that keeps a paused capacity from becoming an outage. Start with the free planner, then bring the scenario to a scoped assessment.

Published 2026-09-25. Service: Cloud & DevOps. Tool: StackSpend Lab. All insights.