Skip to content

Cost Optimization Strategies

This page describes optional strategies to reduce infrastructure cost at scale. Each optimization has a clear trigger point tied to user count — apply them incrementally as the platform grows, not prematurely. For baseline cost components, see Infrastructure; for tiered estimates, see Projections & Profitability.

When to apply each optimization

Optimization Trigger Estimated Savings
Reserved Instances (baseline) 1K users (always-on instances are predictable) ~37% on 2–3 instances
Reserved Instances (peak) 5K users ~$680/month
Graviton migration (r7g/m7g) ✅ Deployed ~20% (already realized)
Spot instances (mixed) 5K+ users (non-critical burst capacity) Up to 65% on burst workers
3-year RI + Savings Plans 50K users (long-term commitment justified) ~57% on all EC2

Reserved Instances / Savings Plans

Commitment Discount Impact at 5K Users
On-Demand 0% Baseline
1-year RI ~37% Save ~$680/month
3-year RI ~57% Save ~$1,045/month

Recommendation

Start Reserved Instances at 1K users for always-on baseline instances. At 5K+ users, 1-year RI on peak instances saves ~$680/month.


Graviton Migration (Deployed)

Current (Graviton) Intel Equivalent Savings
r7g.xlarge ($0.258/hr) r6i.xlarge (~$0.320/hr) ~20%
m7g.large ($0.102/hr) m6i.large (~$0.122/hr) ~16%

Spot Instances (Mixed Strategy)

  • Spot r7g.xlarge: ~$0.044/hr (65% savings)
  • Mixed strategy: 50% On-Demand + 50% Spot
  • On-Demand base for guaranteed capacity

Full cost narrative: Summary & PricingCompetitive Analysis.