Microsoft gives money back through the same three doors as AWS - SLA credits, startup credits, billing recovery - but the mechanics are enterprise-shaped: licensing channels, subscription scopes, and configuration-dependent targets. This playbook walks each door.
Credit types
SLA service credits. Azure consolidates its commitments in the SLA for Microsoft Online Services: per-service uptime targets with tiered credits of 10%, 25%, and 100% of the affected service's monthly charges. Like AWS, credits are claim-only.
| Monthly uptime | Service credit |
|---|---|
| 99% – 99.99% | 10% |
| 95% – 99% | 25% |
| below 95% | 100% |
Applies to the affected service’s monthly charges. Claim required within ~60 days. Max credit 100%. Last verified 2026-07-01. Full per-service tables: SLA credit calculator.
Startup credits. Microsoft for Startups Founders Hub is the most accessible major program - no VC affiliation required. Credits unlock in stages up to $150,000 of Azure, alongside GitHub Enterprise, Microsoft 365, and Azure OpenAI credits. Separate negotiated credits (MACC-related incentives, migration funding) exist for enterprises through account teams.
Billing-error recovery. Azure invoices leak in enterprise-specific ways: reservations bought for architectures that changed, Hybrid Benefit rights owned but not applied, marketplace double-billing, and dev/test subscriptions missing their discounted rates. Audits typically surface 1–3% of annual spend.
Eligibility
- SLA credits: any subscription whose measured uptime fell below the target for its exact configuration. Know your target before claiming - zone-redundant VMs (99.99%), availability sets (99.95%), single premium-SSD instances (99.9%).
- Founders Hub: early-stage software startups; progression through levels is gated on usage and milestones rather than fundraising.
- Billing disputes: any account; EA customers should route through their licensing partner - disputes there move faster than portal tickets.
The claim process, step by step
- Confirm the breach for your configuration's target - the Azure SLA calculator does the tier math instantly.
- Collect evidence: Azure Service Health incident IDs (Portal → Service Health → History), Resource Health timelines for affected resources, your monitoring, subscription and resource IDs.
- File in the right channel: portal support request for pay-as-you-go/CSP; your Microsoft licensing channel for EA. Deadline: end of the second month after the billing month of the incident.
- State the figures: SLA name, subscription, billing month, measured uptime, tier, requested amount against the affected service's charges.
- Reconcile: the credit lands on a subsequent invoice for that subscription; on EAs verify it in the reconciliation file, where credits are easiest to lose.
For Founders Hub: apply before the migration or launch that will burn the credits - they never apply retroactively.
Common mistakes
- Claiming against the wrong SLA target. A single-instance VM claim citing 99.99% gets rejected on configuration grounds even when downtime was real.
- Missing the two-month window - especially common when the incident happened to a subsidiary subscription nobody reconciles monthly. Azure outage alerts put the deadline on the calendar automatically.
- Forgetting Entra ID. Identity outages break everything but claims are filed per affected service - teams claim the app layer and forget the 99.99% Entra ID SLA itself.
- Letting reservations idle after re-architecture. Exchange or refund them (within program rules) instead of eating the waste.
- One-time audits. The leaks recur; schedule quarterly or automate.