Reserved Instances vs Savings Plans Portfolio in 2026

Commitment portfolio is the highest-leverage FinOps decision. Get it wrong by 10% and you waste 5-figures monthly.

Why portfolio matters

Wrong portfolio: locked into capacity you do not use; or paying on-demand for capacity you would commit to.

Right portfolio: 30%+ savings vs all on-demand without lock-in pain.

The 60/30/10 model

• 60% covered by 1-year Compute Savings Plans (flexibility).
• 30% covered by 1-year EC2 Savings Plans (deeper discount).
• 10% reserved for stable known workloads.

Quarterly rebalance

Quarterly: model actual usage; rebalance commitments expiring next quarter.

Most teams find 2-3 percentage points of saving from disciplined rebalance.

Hedging instrument selection

Match instrument to workload: stable + identical = RI; family-stable = EC2 SP; flex needed = Compute SP.

Mixing wrong-instrument-to-workload leaves money on the table.

Antipatterns

• 100% on-demand. 30% savings unrealized.
• 100% RIs. Locked into yesterday.
• No quarterly rebalance. Drift.

What to do this week

Three moves. (1) Apply this lever to your highest-spend workload. (2) Measure the dollar impact for one month. (3) Roll the practice out to the next two services if the savings hold.