AWS · Amazon Web Services

What Is Amazon EC2? Instances, Pricing & Use Cases (2026)

Last verified: June 17, 2026  ·  Format: Breakdown

Amazon EC2 (Elastic Compute Cloud) is secure, resizable compute capacity in the cloud: virtual servers you can launch in minutes, control down to the operating system, and turn off when you are done. If you have been asking what it is in practical terms, picture renting a computer by the hour or the second in Amazon's data centers instead of buying and racking your own hardware. You pick how much processor, memory, and storage you need, run your workload, and pay only while the server is on. It is the flagship compute service inside AWS, and one of the most widely used building blocks in the cloud.

This breakdown is plain and practical. We start with what the service actually is, walk through instances and the key concepts you will meet first, explain the purchase options that drive cost, show how the free tier works as an on-ramp, and close with when this service is the right tool for the job. Service and pricing details below are drawn from AWS documentation and were checked on June 17, 2026; always confirm current figures on the AWS site before you commit.

What Amazon EC2 Is

AWS describes Elastic Compute Cloud as secure, resizable compute capacity in the cloud. In plain terms, it is a service that hands you virtual servers on demand: you choose a configuration, launch it, and you get a running machine you administer like any other server, except the hardware underneath is Amazon's responsibility rather than yours. When you no longer need it, you stop or terminate it and the meter stops.

The headline is choice. AWS offers over 1,000 instances and a choice of Intel, AMD, and Arm-based AWS Graviton processors, so you can match the machine to the work, whether that is a small web server, a memory-heavy database, or a GPU box for machine learning. Underneath, the AWS Nitro System provides built-in security at the hardware level, and the service carries a 99.99% availability service-level agreement. The point of all that variety is that whatever shape of compute you need, a close fit already exists.

Two ideas make this service what it is. First, it is genuinely resizable: you scale capacity up for a spike and back down afterward, rather than provisioning for a worst case you rarely hit. Second, it is metered, so cost tracks usage. If you want the standards-body framing behind those ideas, our what is cloud computing guide covers the model this service is built on, and the what is AWS overview places it alongside the other core services such as storage and serverless.

EC2 Instances and Key Concepts

The unit you actually work with is the instance: a single virtual server, defined by its instance type. The type sets the balance of processor, memory, storage, and networking, and with over 1,000 instances on offer you can pick a small, general-purpose machine for a website or a large, specialized one for high performance computing or model training. Because instances are resizable, a common pattern is to start modest, measure real usage, and move up or out only when the workload proves it needs to.

A handful of supporting concepts make the picture complete:

• Regions and Availability Zones: instances are deployed across geographic Regions, and within a Region across Availability Zones, which are clusters of isolated data centers. Spreading instances across zones is how you build for resilience so a problem in one zone does not take everything down.
• Storage with Amazon EBS: Amazon Elastic Block Store provides block storage that attaches to an instance like a disk, and it is billed per-second with a 60-second minimum, the same metering model as the compute itself.
• Images, firewalls, and scaling: Amazon Machine Images (AMIs), security groups, and Auto Scaling are standard EC2 features that define what an instance boots from, what network traffic it allows, and how a fleet grows or shrinks with demand. The specifics are best read in the AWS EC2 documentation, which carries the current detail.

You do not need to master all of these on day one. Launch a single instance, attach storage, open only the ports you need, and you have a working server. The rest of the concepts come into play as you move from one machine to a resilient, multi-zone deployment.

EC2 Pricing Models

Cost is central to understanding EC2, because the purchase option you choose can change your bill dramatically. All options share the same billing mechanics: per-second billing with a 60-second minimum, so a short-lived instance is charged in seconds rather than rounded up to a full hour. On top of that base, AWS offers several ways to buy the same compute.

The trade-off across these Amazon EC2 options is commitment and flexibility against price. On-Demand costs the most per unit but ties you to nothing. Savings Plans trade a 1- or 3-year commitment for up to 72% off, which suits workloads you can forecast. Spot Instances go furthest on price, up to 90% off spare capacity, but AWS can reclaim them with little notice, so they fit fault-tolerant work that can absorb an interruption. Reserved Instances and Dedicated Hosts round out the set for stable workloads and for cases where you need to bring existing licenses or meet compliance rules. Beyond these, AWS also offers On-Demand Capacity Reservations and EC2 Capacity Blocks for ML when you need guaranteed capacity for a defined window.

Because pricing and discount ceilings change, treat any specific percentage as a starting point and confirm the current numbers on the Amazon EC2 pricing page before you plan around it. A practical pattern is to start On-Demand while you learn the workload, then shift the predictable baseline onto a Savings Plan and push interruptible jobs to Spot.

The EC2 Free Tier

You can try Amazon EC2 through the AWS Free Tier, which is the natural on-ramp for getting hands-on before you spend anything meaningful. The framing changed recently: as of July 15, 2025, new AWS customers choose a free plan, available for 6 months, or a paid plan, and receive up to $200 in AWS Free Tier credits that are usable within 12 months. In practice that lets you launch a real instance, attach storage, and learn by doing rather than reading.

The credit model is worth understanding before you rely on it. Credits are a finite allowance, not an unlimited free machine, and an instance left running can quietly consume them. Set a budget and watch usage from the first day, the same discipline that keeps any metered cloud bill predictable. Because free-tier terms and credit amounts change, confirm the current offer on the AWS site before you plan around it.

When to Use Amazon EC2

We have covered what the service is and how it is priced; the last question is when it is the right tool. EC2 shines when you want control over the server itself, the operating system, the runtime, and the configuration, and when your workload does not fit a fully managed or serverless model. AWS calls out several common scenarios, and the personas below show where each one lands.

Honest Trade-offs

Every cloud platform has trade-offs, and EC2 is no exception. EC2 is a sensible default when you need control of the server, and the points below are not reasons to avoid it. They are reasons to adopt it with clear eyes.

Frequently Asked Questions