This post will save you thousands of headaches and will make your business operations stand out from everyone else who follows the same, sometimes irrational, argument that you need the cloud. No, you don't need the cloud. In fact, there are very few, rare scenarios where you really need the cloud. Those are reserved for companies making at least $50M, and even then, you won’t need it for long.
How did we get to this point? To be completely transparent, I have been in the industry for the last five years and recently decided to go all-in on my projects. I'm relatively young (23) for this industry, but that doesn't change the fact that I've seen, with my own eyes, the harm caused by the wrong mindset, questionable marketing, and the current startup culture. This irrationality has to stop.
The big three cloud providers (AWS/GCP/Azure) have done an incredible marketing job over the last decade to make you think you need them. Now, in 2025, they’re considered the standard for startups to host and run their businesses. They’re capitalizing on a new generation of developers and entrepreneurs who don’t fully understand how the internet actually works. If you did, you'd quickly realize you’re overpaying for compute and simple server tasks.
I’ve worked with dozens of businesses and talked to entrepreneurs who think this is the norm and that there’s no other way to test a simple MVP. You’ve been brainwashed with promises of infinite scalability because you believe your business will take off overnight with a sudden spike in traffic (and I hope it does). But in reality, the price of a simple, powerful 8 vCPU, 32GB RAM, 240GB NVMe VPS that can run applications for years serving millions of users is all you need for MVPs and critical production applications.
When you run a business with investor capital, saving costs is your lowest priority because you have to "get into the market fast," "product-market fit," and "make investors happy." Honestly, when it’s not your own money on the line, you couldn’t care less. It makes sense to pay for all the unnecessary products the cloud offers, after signing a contract that locks you in with one vendor for the rest of your business life. Again, cloud providers are capitalizing on this mindset and the money-driven decisions of VCs. It’s the perfect formula for burning money: a non-technical founder, a VC who only cares about financial returns, and big tech companies monetizing their infrastructure. The result? Poor business decisions.
When executives ask developers how to build infrastructure, do they hear…?
"Let’s spin up a VPS, use Coolify or Kubero to display a UI for Docker deployments in our VPN protected by WireGuard, run a couple of instances for PostHog and Plausible for analytics, and use a GitHub Action for automatic deployments on the main branch private repo via webhook. We can maintain all of this on a $25/month VPS for years."
Or do they hear:
"We have to choose between AWS, GCP, or Azure, but I’ll need a couple of weeks to calculate the costs because, you know, cloud pricing is unpredictable. But hey, we get infinite scalability!"
When you talk to these engineers, it feels like you are talking to salesmen that don't realize they are paying for open source wrappers. They’ve dedicated a lot of time to learning these purposefully complex systems with unique names, only to realize all you need is Linux and Docker. It’s mind-blowing for me to see people with this knowledge recommending the cloud when they know deep down it’s expensive. With the same knowledge, you could run this for 10x less. It might sound funny, but the only people who defend irrational arguments while ignoring objective points are cults. And it’s not ironic that “cloud people” online behave like obsessive fans. Just an observation, not a general rule.
This is a business. The priority is always high profits and low operational costs. That’s the only rule that matters. If you know this and still decide to go with the cloud, you’re either not a businessperson (because it’s not your money on the line) or your calculations need serious review when comparing self-hosting to the cloud.
You might think specialized knowledge is needed to manage an infrastructure, and I don’t have it. First, I do have a solid grasp of sysadmin, Linux, Docker, and cybersecurity, enough to understand the cloud offers the exact same things, just with different names and a UI. And let’s be honest, you’ll still need to touch the “scary console”, even with the cloud. Don’t let these companies implant a software development paradigm that only benefits them. Applying simple business common sense, technical skills, and optimization, the cloud is unnecessarily expensive for most cases.
This applies to heavy workloads like AI/ML or massive elasticity with load balancers. I’m not saying the cloud isn’t necessary, but for 90% of cases, it’s not suitable. The remaining 10%? Big companies that, for some reason or lack of DevOps engineers, don’t want to keep infrastructure in-house, and for whom funding isn’t a problem. The only reasonable arguments for the cloud are "timing the market" or compliance with certain industry regulations. Once a business is comfortable orchestrating multiple instances, the cloud’s “benefits” shrink drastically.
The cloud’s unpredictable pricing and lack of transparency should be enough reason to reconsider and stay away from it for as long as possible, no matter what stage your business is at. If you’re indie or small business definitely stay away from them.