As I promised in my previous newsletter, over the next few weeks, we will be going over specific strategies for how to reduce your company’s cloud, software, AI, and other IT expenses. This week, we’ll be focused on AWS cost savings opportunities.

I am going to repeat my previous disclaimer: not all of these cost savings recommendations will be applicable to your company’s particular cloud environment. That being said, being aware of all potential cost savings opportunities can help when planning future infrastructure changes.

Many engineers feel stalled because the role itself has not evolved. The work looks the same, but the market has moved.

Senior engineering in 2026 demands ownership, faster judgment, and comfort with ambiguity. If your role is not pushing you there, it may be holding you back.

Last cohort, 15 hiring partners sent 31 representatives to evaluate challengers through 246 live interviews. Gauntlet offers a reset. Apply now.

Cohort 6 starts July 6

_{Must be a US citizen to qualify.}

One of the most common causes of excessive AWS spending is oversized EC2 instances. Many organizations initially provision workloads based on anticipated peak demand and never revisit those sizing decisions. As a result, servers often operate with low CPU and memory utilization while continuing to incur premium costs.

AWS provides tools such as Compute Optimizer and CloudWatch that make it easier to identify underutilized resources. By analyzing utilization metrics and matching workloads to appropriately sized instances, organizations can often reduce compute spending by 20 to 50 percent without impacting application performance.

Not every workload needs the flexibility of on-demand pricing. Production servers, databases, and other long-running applications often have predictable usage patterns that make them ideal candidates for Reserved Instances.

By committing to a one-year or three-year term, organizations can receive significant discounts compared to standard pricing (up to 75%). The downside with Reserved Instances is that you are required to commit to specific types of compute (e.g., a specific region, operating system, and size) and you will be restricted from changing your commitment during the Reserved Instance term.

Instead of committing to a specific instance family with a Reserved Instance, you can instead choose to commit to a consistent level of compute spend over time across multiple services with Savings Plans. Savings Plans have become a popular alternative to traditional Reserved Instances because they provide similar discounts (up to 72%) while offering greater flexibility by automatically applying discounts across eligible services and instance types. For organizations with evolving environments or multiple application teams, Savings Plans often deliver substantial savings without sacrificing operational agility.

Not every workload requires guaranteed availability. Batch processing jobs, data analytics workloads, testing environments, and machine learning training tasks can often tolerate interruptions without impacting business operations.

AWS Spot Instances provide access to unused compute capacity at dramatically reduced prices. Although these instances can be reclaimed by AWS when capacity is needed elsewhere, organizations that design workloads for interruption tolerance can achieve some of the largest cost reductions available in the cloud.

Many organizations focus heavily on production costs while overlooking development and testing environments. These systems frequently remain powered on around the clock despite being used only during business hours.

Automating startup and shutdown schedules can produce immediate savings. A development environment that runs only during working hours may reduce infrastructure costs by more than 60% compared to operating continuously. Simple automation tools can ensure resources are available when needed while eliminating unnecessary overnight and weekend expenses.

Cloud infrastructure should adapt to demand rather than remaining fixed at peak capacity. Yet many organizations continue operating workloads with static resource allocations that generate unnecessary costs during periods of low activity.

AWS Auto Scaling allows resources to expand during busy periods and contract when demand decreases. This ensures applications remain responsive while preventing organizations from paying for unused capacity. Over time, dynamic scaling can significantly improve both cost efficiency and operational performance.

S3 Storage costs may seem small on a monthly basis, but they accumulate quickly as organizations retain increasing amounts of data. Many businesses store all files in premium storage tiers regardless of how frequently the data is accessed.

AWS offers multiple storage classes designed for different access patterns. By implementing lifecycle policies that automatically move older data into lower-cost tiers such as Glacier or Deep Archive, organizations can dramatically reduce storage expenses while maintaining access to important information when needed. Note that data stored in cold storage tiers will take longer to access and transfer, which may not make them suitable in all production environments.

Data transfer costs are one of the most misunderstood components of AWS billing. While compute and storage expenses receive significant attention, networking charges can quietly become a substantial contributor to overall cloud spend.

Organizations should carefully review how data moves between regions, availability zones, and external networks. Reducing unnecessary transfers, consolidating workloads, and optimizing content delivery strategies can help control these often-overlooked expenses.

Cloud environments naturally accumulate unused resources over time. Old snapshots, unattached EBS volumes, forgotten load balancers, and unused Elastic IP addresses often remain active long after projects have ended.

Because these resources are rarely visible in day-to-day operations, they can generate ongoing charges without providing any business value. Conducting regular cleanup reviews helps eliminate waste and ensures cloud spending aligns with actual usage.

Databases are often among the most expensive services in an AWS environment. Like virtual machines, database instances are frequently oversized or configured with more storage and performance than necessary.

Regular reviews of database utilization can uncover opportunities to reduce costs through rightsizing, storage optimization, or reserved capacity purchases. Organizations should also evaluate serverless database options where appropriate, especially for workloads with variable demand patterns.

Traditional virtual machines frequently leave computing resources underutilized. Containers allow multiple applications to share infrastructure more efficiently, improving overall resource utilization.

Services such as Amazon ECS and Amazon EKS make it easier to deploy and scale containerized applications. Beyond the operational benefits, containerization often reduces infrastructure costs by enabling organizations to run more workloads on fewer underlying resources.

Cost optimization becomes difficult when organizations lack visibility into resource ownership. Without consistent tagging standards, it can be nearly impossible to determine which teams, applications, or departments are responsible for cloud spending.

A strong tagging strategy improves accountability and enables more accurate cost reporting. When teams understand exactly what they are spending and why, they are more likely to identify and eliminate unnecessary expenses.

The AWS Enterprise Discount Program (EDP) is a contractual agreement between an organization and Amazon Web Services that provides discounted pricing in exchange for a long-term spending commitment. Designed primarily for large enterprises and organizations with significant cloud consumption (typically at least $1M per year), EDP agreements typically span one to five years and establish a minimum annual or total spend commitment across eligible AWS services.

The primary benefit of an EDP is reduced AWS costs through negotiated discounts that go beyond standard pricing programs such as Savings Plans or Reserved Instances. As organizations increase their cloud spending commitments, they may qualify for larger discounts, making EDPs particularly attractive for businesses undergoing large-scale cloud migrations, digital transformation initiatives, or rapid growth.

In addition to pricing advantages, EDP agreements may include access to additional AWS resources, such as dedicated account support, funding programs, training credits, migration assistance, and strategic guidance from AWS specialists. These benefits can help organizations accelerate cloud adoption while maximizing the value of their AWS investments.

Note that entering into an EDP agreement will require your organization to sign up for Enterprise Support, which will cost at least $5k/month or 10% of annual monthly bill up to $150k, whichever is greater. The additional cost of paying for Enterprise Support will need to be factored in when calculating potential cost savings stemming from signing an EDP agreement.

Graviton Processors are a class of processors that can be utilized for your EC2 instances. Graviton instances can be up to 20% cheaper than equivalent Intel or AMD instances. They also typically deliver better performance per dollar on workloads optimized for ARM.

Reducing AWS costs is rarely about finding a single dramatic change. Instead, the greatest savings come from consistently applying a combination of technical best practices, financial discipline, and operational governance.

Organizations that proactively manage their cloud environments can often reduce spending significantly while improving efficiency and reliability. By focusing on these strategies, businesses can gain greater control over their AWS costs and ensure their cloud investments continue delivering maximum value.

Next week we will focus on one of the hottest topics in Enterprise IT: reining in AI costs.