Massive cloud investment continues worldwide, with Gartner predicting public cloud spending to reach an eye-popping $1 trillion by 2027. This number is growing significantly as companies invest more in generative AI, as GenAI initiatives require a lot of cloud capacity.
And yet, many organizations still struggle to maximize the value of their cloud investments. Cloud waste is a rampant problem; it’s estimated that anywhere from 28-35% of cloud spend is wasted. It’s little wonder, then, that a recent CloudZero survey found that 72% of respondents said their cloud costs were either “too high” or “way too high.”
How do you get the most bang for your buck? It starts with taking a different approach to how you think about and use the cloud.
The cloud waste problem
The right mindset involves veering away from the “lift and shift” mentality of just taking existing resources and moving them into the cloud. Cloud waste stems largely from this antiquated mindset, which treats cloud infrastructure like traditional infrastructure.
The consumption and management of cloud infrastructure has little in common with traditional infrastructure. Before the cloud, companies invested heavily in data centers and servers, spending outsized sums of money on the infrastructure they thought they’d need to process the demand they expected to generate. The process was: Product teams proposed some innovation, predicted demand, and made formal requests to IT procurement teams for the infrastructure they expected to need. The procurement team could approve, deny, or modify the request, and months later, the product teams might have the infrastructure they’d need to execute the innovation.
Companies often bought more infrastructure than they ended up using, and found themselves sitting on servers that weren’t generating any value. Virtualization promised to even this balance, but over-provisioning and under-utilization continued to be a challenge. And while the cloud has introduced endless possibilities through a diverse set of infrastructure, database, and platform services and a consumption-driven utility model, many companies still manage it like a collection of physical virtual machines.
Procurement and finance teams used to be involved in every infrastructure purchase. Now, in the cloud, infrastructure consumption happens instantly, whenever an engineer spins up a new cloud resource or writes a line of code that consumes those resources. The purchase moment has changed entirely: In the cloud, every engineering (building) decision is a buying decision. Engineers — not finance leaders or centralized IT teams — are directly spending the company’s technology budget.
So, when companies pin cloud costs on finance teams or centralized IT teams, they miss the mark. Engineers make building decisions based on engineering expertise — expertise that other teams don’t have. Finance teams can make bulk purchases or optimized committed use discounts, but you do not want them distinguishing between the use of a m7g.2xlarge and a m7gd.metal. IT teams are great at finding underutilized resources, but they are not in the best position to understand if the code running on a highly utilized resource is healthy or not. In the cloud, buying better only gets you so far.
For a long time, engineers have lacked the financial insight to make cost-efficient building decisions in the cloud, leading to a torrent of cloud waste annually. A recent survey by CloudZero found that companies that implement formal cloud cost management programs tend to reduce their annual cloud spending by 20-30%. Given that 61% of companies don’t have formalized programs, this means that when cloud spending hits $1 trillion in 2027, as much as $122–183 billion of that could be wasted.
This needs to change. Companies need to realize that cloud infrastructure is entirely different from traditional infrastructure, and that cloud cost management requires a completely new approach. We need to shift away from buying better to building better: equipping engineers to take ownership of their own cloud costs, and, as Amazon CTO Werner Vogels put it in The Frugal Architect, “make cost a non-functional requirement” of great software.
Time to build better in the cloud vs. buying better
Building better is an engineering philosophy rather than a financial paradigm. “Building” refers to every architecture, coding, or operations decision engineers make in the process of developing a product and bringing it to market.
Until recently, there hasn’t been a way to grasp the true cost of such decisions, and organizations weren’t very invested in finding out. The mindset of buying better comes from a reactive desire to reduce costs, whereas the mindset of building better is all about developing and running efficient software.
Benefits of building better
Engaged engineers. The data suggests that when engineers are equipped to manage their own costs, they do — and that companies perform better. In that same survey, 81% of companies said cloud costs are “about where they should be” when engineers had some level of ownership over cloud costs. Focusing on building better means focusing squarely on engineering engagement: giving engineers relevant, timely data about their cloud infrastructure costs, and making it easy to track efficiency gains.
Improved finance-engineering relations. When companies focus on building better, it allows finance and engineering teams to focus on their respective specialties. Engineers weigh the factors that go into well-architected software; finance teams get regular, detailed reports about the cost efficiency of that software. The friction between the teams is reduced, and overall productivity improves.
Unit economic clarity. Giving engineers meaningful cost data means ingesting all spend data (beyond just the hyperscalers to include platform services, database services, observability tools, etc.) and allocating it in a framework that mirrors the company’s business. Such robust allocation yields the material for cloud unit economics: assessing profitable and unprofitable products, features, and customers, understanding fixed versus variable costs and the relationships to margins, and refining your GTM strategy based on this data. Cloud unit economics is the holy grail of cloud financial operations (FinOps) — and the mark of a truly cloud efficient organization.
It’s time to build better
More and more organizations feel that they’re getting too little return on their cloud investments. By switching from a buying better to a building better approach, organizations gauge their approach to the true nature of the cloud, producing better engineering engagement, improved relations between finance and engineering teams, and stronger unit economics.