Always-On PTUs for Seasonal or Cyclical Azure OpenAI Workloads

When PTUs remain fully provisioned 24/7, organizations incur continuous fixed cost even during extended idle time. Although Azure does not offer native PTU scheduling, teams can use automation to provision and deprovision PTUs based on predictable cycles. This allows them to retain performance during peak windows while reducing cost during low-activity periods.

On hourly billing, PTUs are charged for as long as the deployment exists, prorated to the minute (deployments cannot be paused), while Azure Reservations cover a fixed PTU quantity at a discount for sustained use. If workloads operate on weekly, monthly, or seasonal cycles, keeping PTUs active outside of demand windows results in paying for idle dedicated capacity.

• Identify PTU-backed workloads with recurring or predictable activity patterns
• Review utilization trends showing regular periods of inactivity despite continuous PTU allocation
• Evaluate whether workload timing aligns with business cycles, reporting schedules, or seasonal events
• Assess whether PTUs remain active during nights, weekends, or large idle blocks

• Use automation to scale PTUs up and down according to expected workload schedules
• Reserve scheduling for short, predictable demand windows: Microsoft advises against scaling production deployments up and down on hourly billing, because a right-sized reservation is typically cheaper for sustained production use and scaling down releases capacity that may not be available when you scale back up
• Provision PTUs only during windows where predictable throughput and low latency are required
• Establish scheduling policies for cyclical workloads to prevent unnecessary continuous provisioning
• Regularly re-evaluate patterns to ensure scheduling aligns with evolving business cycles

• Provisioned throughput billing and cost management - Microsoft Foundrylearn.microsoft.com
• Provisioned throughput for Foundry Models - Microsoft Foundrylearn.microsoft.com
• Azure OpenAI Service - Pricingazure.microsoft.com