5 learnings to stop cloud overspend
August 27, 2025
3 min read
Does paying your cloud bill feel like throwing money out the window? Tim Banks and Justin Garrison recently spoke about this in our webinar “Optimization alone won’t save your Kubernetes budget,” where they dug into the value proposition of cloud, how it has changed over the years, and what you can really do to put a stop to cloud overspend.
Here are five of our favorite takeaways from the webinar.
1. Cloud lets us pile up environments like clutter in a garage
At first glance, cloud looks cheap. It IS cheap if you only use a little. Of course, that rarely happens. There’s always more compute, more storage, and then companies end up having no idea what they’re storing. In the end, cloud has made us reckless, encouraging unchecked growth as teams spin up resources without tracking usage or value.
Like a cluttered garage, infrastructure piles up, often with redundant environments no one notices, and the cost becomes much higher than ever intended.
2. You don’t need more stuff to move fast
Urgency pushes teams to focus on getting things done now, and the easiest way is to add various tools and platforms. When you’re focused on accelerating, optimizing, and automating in the short term, it seems like an obvious solution.
As a result, teams end up believing that fast equates with more. More layers, more tools, more things. This isn’t necessarily true. What most teams need is to simplify. Kubernetes can be bloated as it is, and there’s no reason to further complicate your infrastructure with stuff.
3. Efficiency falls to the side because it isn’t rewarded
Would you do work that your company doesn’t really notice?
Inefficiency is hardly considered an emergency. Organizations prefer to celebrate the new and additional revenue, overlooking the power of cutting costs through efficiency. This leaves teams disincentivized to closely examine their infrastructure efficiency, making it unlikely there will be a close examination of all the radically different options or missed opportunities.
Over time, many have forgotten how to calculate spend, define requirements, or understand what infrastructure really costs.
4. It’s time we relearn infrastructure
As cloud-native replaced traditional ops roles, teams lost practical skills like running a data center or managing physical infrastructure. Now, teams need to relearn the tipping points of infrastructure and understand when to build, when to buy, and when to rethink the entire approach. This should include refreshing market knowledge to understand what is cheaper or more efficient today.
5. Start small, plan smart
Greenfield projects are the best time to reassess assumptions. You can start small by running a service on-prem to see what’s needed. You’ll quickly become familiar with topics like storage, network, compute, backup, caching. Plus, you’ll start assembling quotes and other industry information that you’ll need should you decide to continue moving your infrastructure.
A final work: Question the “defaults.” Cloud isn’t always cheaper. You don’t always need the maximum performance you think you do. Kubernetes isn’t always right-sized. The only way to know something for sure is to come up with a hypothesis, and start testing! Want to watch the full webinar? Check it out below and get all the details I couldn’t add to this summary. There are plenty more insights and tips to help you stop overspending on cloud.
Steve Francis