Skip to main content
In this lesson, we dive into AWS EC2 pricing and explore six cost models designed to optimize your cloud spend. By understanding how each option works, you can align your infrastructure costs with your application’s growth and usage patterns.

​
Cloud Growth Phases

Every organization’s cloud journey typically moves through three phases:
  • Start-up Phase
    Experimentation with limited users, resources created and terminated on demand, pay-per-use billing.
  • Steady Phase
    Achieved product-market fit, predictable 24/7 usage, consistent resource requirements.
  • Growth Phase
    Fluctuating hourly demand, advanced analytics and batch processing, high user volume.
The image illustrates the phases of cloud growth: Start-up Phase, Steady Phase, and Growth Phase, each with specific characteristics like on-demand services, regular traffic, and complex functionalities.
As your application evolves, selecting the right EC2 pricing model is key to achieving cost efficiency without sacrificing performance.

​
Why EC2 Cost Management Matters

Amazon EC2 often represents a significant portion of your AWS bill. Gaining clarity on EC2’s billing mechanisms empowers you to:
  • Identify idle or underutilized resources
  • Apply the most cost-effective pricing option
  • Forecast and control cloud expenditure
The image is about "EC2 Cost Management" and features a pie chart highlighting EC2, along with icons representing AWS Cloud and cost-saving ideas.
AWS provides six pricing models for EC2 instances:
  • On-Demand Instances
  • Spot Instances
  • Savings Plans
  • Reserved Instances
  • Dedicated Hosts
  • On-Demand Capacity Reservations
The image outlines the AWS EC2 pricing model, listing six options: On-Demand Price, Spot, Savings Plans, Reserved Instances, Dedicated Hosts, and On-Demand Capacity Reservation.
To compare these options at a glance, refer to the table below:
Pricing ModelIdeal Use CaseCommitmentPotential Savings
On-DemandSpiky workloads, dev/testNoneNone
Spot InstancesBatch jobs, fault-tolerant appsNoneUp to 90%
Savings PlansConsistent compute spend1–3 yearsUp to 66% (Compute) / 72% (EC2)
Reserved InstancesPredictable, long-term workloads1–3 yearsUp to 72%
Dedicated HostsBYOL, compliance, regulatory requirementsNoneAdditional RI/Savings Plan discounts
On-Demand Capacity ReservationsBusiness-critical & high-availability needsOn-demandNone

​
On-Demand Instances

On-Demand Instances let you pay by the second with no long-term commitments. AWS allocates capacity immediately upon launch, and you’re charged only for the time the instance runs. Features:
  • No upfront fees
  • Per-second billing
  • Full flexibility to launch and terminate
Use cases:
  • Development and testing environments
  • Spiky or unpredictable traffic patterns
  • Applications with short-term or intermittent workloads

​
Spot Instances

Spot Instances provide access to spare EC2 capacity at discounts of up to 90% off On-Demand prices. AWS may reclaim Spot Instances with a two-minute warning when capacity is needed elsewhere.
The image illustrates the concept of AWS Spot Instances, showing a diagram of cloud resources with price matching and on-demand requests, alongside icons representing features and use cases.
Key points:
  • Variable pricing based on supply and demand
  • Two-minute interruption notice
  • No upfront commitment
Use cases:
  • Stateless, fault-tolerant workloads
  • Batch processing and big data analytics
  • Flexible start and end times
Spot Instances are ideal for workloads that can gracefully handle interruptions. Consider integrating checkpointing or auto-scaling to reduce impact.

​
Savings Plans

Savings Plans offer significant discounts in exchange for a constant hourly spend commitment over a one- or three-year term.
The image illustrates AWS Saving Plans, showing a comparison between on-demand and discounted prices, and highlighting Compute and EC2 Instance Savings Plans.
Two plan types:
  1. Compute Savings Plan
    • Up to 66% savings
    • Applies to any EC2 instance, regardless of family, region, OS, or tenancy
  2. EC2 Instance Savings Plan
    • Up to 72% savings
    • Locked to a specific instance family in one region
    • Flexibility to change instance size, OS, and tenancy within that family
Savings Plans automatically apply your commitment to the lowest-cost usage across eligible instance types.

​
Reserved Instances

Reserved Instances (RIs) are an older commitment model offering deep discounts in exchange for a one- or three-year term.
The image is an infographic about AWS Reserved Instances, showing a comparison between Standard and Convertible Reserved Instances with savings percentages and flexibility options.
Types of RIs:
  • Standard Reserved Instances
    • Up to 72% off On-Demand pricing
    • Modify availability zone, instance size, and network platform
  • Convertible Reserved Instances
    • Up to 66% off On-Demand pricing
    • Exchange one RI for another of equal or greater value
    • Change instance families, OS, and tenancy

​
Dedicated Hosts

Dedicated Hosts allocate a physical server solely for your use, enabling you to bring existing software licenses and meet strict compliance requirements.
The image is an infographic about AWS Cloud dedicated hosts, highlighting features like BYOL and reserved instance savings plans, and detailing specifications for the c7g instance family.
Key features:
  • Full control over sockets and cores
  • Ideal for BYOL, compliance, and regulatory mandates
  • Combine with RIs and Savings Plans for extra discounts
Example c7g host capacity:
  • 192 medium instances
  • 96 large instances
  • 48 xlarge instances
  • 24 2xlarge instances

​
On-Demand Capacity Reservations

On-Demand Capacity Reservations let you reserve compute capacity in a specific Availability Zone, ensuring capacity is available when you need it.
The image is an illustration of AWS On-Demand Capacity Reservation, highlighting features such as capacity assurance, regulatory HA requirements, disaster recovery, and support for RI and savings plans.
Use cases:
  • Business-critical applications requiring guaranteed capacity
  • Regulatory or compliance mandates for high availability
  • Disaster recovery and failover scenarios
You are billed for your reserved capacity whether or not you run instances. Apply RIs or Savings Plans to offset reservation costs.
With this overview of AWS EC2 pricing models, you can match your workloads to the most cost-effective option and optimize your cloud investment.

Watch Video