Amazon Elastic Compute Cloud (Amazon EC2) is AWS’s service for renting resizable virtual servers in Amazon data centers. You launch an instance from an Amazon Machine Image (AMI), select capacity and networking settings, run your software, and pay according to the pricing model and resources you use. EC2 is virtual capacity—not a physical server that you purchase.
What is Amazon EC2?
Amazon Web Services describes EC2 as a web service that provides “resizable computing capacity—literally, servers in Amazon’s data centers—that you use to build and host your software systems.” In practical terms, EC2 lets you create virtual machines on demand instead of buying and operating your own server hardware.
You can increase or decrease the number or size of instances as demand changes. An instance can host a website, API, application server, development environment, batch job, or other software that needs a computer in the AWS Cloud.
What is an EC2 instance?
An EC2 instance is a virtual server running on AWS infrastructure. Its operating system, initial software, and configuration come from the image you choose at launch.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →The AMI determines the starting image
An Amazon Machine Image (AMI) is the template used to launch an instance. It supplies the operating system and can include an application server and applications. Choosing a Linux- or Windows-based AMI, for example, changes the software environment that starts on the instance.
Virtual capacity, not owned hardware
AWS runs the underlying physical data-center infrastructure. You select virtual capacity and operate the guest system and applications, while AWS manages the physical host and service infrastructure according to its cloud service responsibilities.
Rank #2
How EC2 capacity is purchased
AWS offers several models because price, flexibility, interruption tolerance, capacity guarantees, and hardware or licensing requirements differ by workload.
| Option | How it works | Best fit | Main trade-off |
|---|---|---|---|
| On-Demand Instances | Pay for usage without a long-term commitment. | Variable, short-lived, or exploratory workloads. | Usually the least discounted option; there is no commitment-based price reduction. |
| Savings Plans | Commit to a specified level of eligible hourly usage for one or three years in exchange for discounted pricing on eligible use. | Workloads with predictable, continuing spend but some flexibility in instance usage. | You owe the committed usage level for the term, even if your actual demand falls below it. |
| Reserved Instances | Commit for one or three years to a matching instance configuration. | Stable workloads whose instance family, region, and configuration are well understood. | Less flexible when your required configuration changes. |
| Spot Instances | Use spare EC2 capacity at a discount. | Interruptible batch processing, distributed jobs, test environments, and other fault-tolerant work. | AWS can interrupt the capacity, so the workload must recover or restart safely. |
| Dedicated Hosts | Reserve a physical server dedicated to your organization’s instances. | Requirements involving dedicated hardware, visibility of host placement, or certain license terms. | Higher cost and more operational planning than shared infrastructure. |
| Dedicated Instances | Run instances on single-tenant hardware. | Workloads that require hardware tenancy separation without managing an entire dedicated host. | Less control than a Dedicated Host and a different pricing profile from shared tenancy. |
| Capacity Reservations | Reserve EC2 capacity in a specified Availability Zone. | Applications that must be able to launch capacity in a particular Availability Zone when needed. | You pay for the reservation while it is active, whether or not you fully use the capacity. |
Choosing among On-Demand, Spot, and Savings Plans
- Choose On-Demand when avoiding a commitment matters more than obtaining a lower rate.
- Choose Spot only when interruption is acceptable and the application can checkpoint, retry, or fail over.
- Choose a Savings Plan when you have a reliable baseline of eligible usage and can commit for one or three years.
AWS advertises Savings Plans at up to 72% below On-Demand prices and Spot capacity at up to 90% below On-Demand prices. These are AWS-published maximum comparisons for the 2026 access year, not guaranteed savings for an individual account; actual results depend on product terms, region, instance use, and utilization.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How much does Amazon EC2 cost?
There is no single EC2 price. Your bill depends on the instance capacity and operating-system image you run, how long resources remain active, the purchasing model, and related AWS resources used by the workload. The same application can have very different costs under On-Demand, a commitment plan, or Spot.
Costs to account for
- Compute usage: the running instance and its selected capacity.
- Commitments or reservations: Savings Plans, Reserved Instances, and Capacity Reservations can create charges or obligations independent of short-term utilization.
- Tenancy: Dedicated Hosts and Dedicated Instances have different pricing from shared tenancy.
- Workload design: scaling out, leaving instances running while idle, and selecting larger capacity all affect usage.
Use the current AWS pricing pages and your account’s billing tools for an estimate. Discount percentages and other prices change over time, and a quoted rate should always be labeled with its region, instance configuration, operating system, pricing model, and date.
Is EC2 free?
EC2 is not universally free. AWS states that Free Tier benefits depend on when the account was created. For accounts created on or after July 15, 2025, the current launch guidance describes certain instance types for six months or until included credits are used. Eligibility, instance types, credits, and other terms must be checked in the account’s current offer.
Even when a Free Tier benefit applies, charges can accrue for running instances and related resources, including when an instance is idle. Stop or terminate resources you no longer need, and monitor billing rather than assuming that an unused application is cost-free.
How an EC2 launch fits together
- Select an AMI: choose the operating system and any included application software.
- Select instance capacity: match virtual CPU, memory, storage, and network requirements to the workload.
- Choose tenancy and purchasing terms: decide whether shared, dedicated, reserved, committed, or interruptible capacity is appropriate.
- Place the instance: select the AWS Region and, where relevant, an Availability Zone or a Capacity Reservation.
- Configure access and workload controls: apply the account’s networking, identity, storage, monitoring, and operating-system settings.
- Run and manage the software: patch the guest operating system, scale capacity as demand changes, and shut down resources that are no longer required.
The exact options shown depend on the selected AMI, instance type, Region, account, and AWS service changes, so confirm the current console and pricing labels before launching production capacity.
EC2 security: who is responsible?
AWS uses a shared responsibility model. AWS protects the infrastructure that runs its services—“security of the cloud.” The customer is responsible for security within their cloud use—“security in the cloud.”
For EC2, customer responsibilities commonly include choosing a trusted AMI, configuring identity and access controls, restricting network exposure, patching and hardening the guest operating system, protecting application data, and monitoring activity. The precise duties depend on the workload and configuration. AWS’s physical facilities and underlying infrastructure remain AWS responsibilities, but that does not secure an incorrectly configured instance or application.
When EC2 is a good fit
- You need control over an operating system and application runtime.
- Workload demand changes and you want capacity that can be resized.
- You are migrating an existing server-based application to the cloud.
- You can design for interruption and use Spot for suitable jobs.
- You need dedicated tenancy or a capacity guarantee in a particular Availability Zone.
EC2 requires more system administration than a fully managed application service: you remain responsible for the guest environment and the software you install. That control is its advantage when your application needs it, but it also makes configuration, patching, monitoring, and cost controls essential.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
