Case against the cloud

Cloud computing has swept the business world, with many companies rushing to migrate their operations to cloud-based solutions or building new projects entirely in the cloud. The promise of scalability, flexibility, and reduced maintenance sounds appealing, but I believe the cloud is often overrated for many projects. It can be an expensive choice that locks businesses into rigid, costly systems. By exploring real-world examples and weighing the risks, I’ll explain why the cloud may not always be the best path—and why owning your infrastructure might save both money and headaches.

The Allure and Risks of Cloud Migration

The cloud offers clear benefits: quick setup, automatic scaling, and no need to manage physical servers. These advantages make it tempting for startups and established businesses alike. However, the cloud comes with a hidden trap—vendor lock-in. Once a company commits to a cloud provider, moving to another platform or back to on-premises infrastructure can be complex and costly. Switching providers often means rewriting code, reconfiguring systems, and migrating data, all of which demand time and expertise. If the chosen cloud solution lacks needed features or its costs skyrocket, businesses can find themselves stuck, unable to pivot without significant expense.

This lock-in risk is not just theoretical. Many companies discover too late that their cloud provider’s limitations or pricing models do not align with their long-term needs. The initial ease of the cloud can turn into a costly burden, especially for projects that do not require the cloud’s full power.

A Case Study: Ahrefs’ Cost Savings

Ahrefs, a company known for its SEO tools, provides a compelling example of the financial impact of avoiding the cloud. In a detailed analysis, Ahrefs compared the costs of running their infrastructure in a rented colocation data centre in Singapore against using Amazon Web Services (AWS), a leading cloud provider. Their findings were striking: by sticking with their own infrastructure, Ahrefs saved approximately 400 million US dollars over two and a half years. This figure continues to grow as they expand their colocation data centre with new hardware.

Ahrefs’ approach shows that owning and managing physical servers can be far more economical than relying on cloud services, especially for businesses with stable, predictable workloads. Their homogenous infrastructure allowed them to optimise costs without sacrificing performance, challenging the assumption that the cloud is always the cheaper option.

37signals: Moving Back to Owned Hardware

Another example comes from 37signals, the creators of Basecamp and Hey. In 2022, they spent 3.2 million US dollars on cloud services, averaging 266,797 US dollars per month. Realising this was unsustainable, they began moving services back to their own hardware in 2023 to cut costs significantly. However, the transition was not simple. They had to develop new tools for managing their infrastructure, rethink their deployment and monitoring strategies, and adapt their continuous integration and delivery pipelines. This added complexity and technical debt, highlighting the operational challenges of leaving the cloud.

Despite these hurdles, 37signals found the cloud’s costs outweighed its benefits, both financially and operationally. Their openness about their 2022 spending and their decision to invest in owned hardware underscores a growing scepticism about the cloud’s one-size-fits-all appeal.

Weighing the Trade-Offs

The experiences of Ahrefs and 37signals reveal a broader truth: the cloud is not always the most economical or practical choice. For projects with consistent workloads or specific performance needs, owning infrastructure can save millions while avoiding lock-in. However, moving away from the cloud requires careful planning. Companies must build robust processes for managing servers, monitoring performance, and ensuring reliability—tasks that the cloud typically handles automatically.

For smaller projects or those with unpredictable scaling needs, the cloud may still make sense. Its ease of use and flexibility can outweigh initial costs for startups or apps in early development. But as businesses grow, they should evaluate whether the cloud’s convenience justifies its price tag and the risk of being tied to a single provider.

A Balanced Approach

Before jumping to the cloud, businesses should assess their project’s needs. Does the app require massive scaling, or will traffic remain steady? Can the team manage physical servers, or is the cloud’s automation worth the cost? By answering these questions, companies can avoid the trap of assuming the cloud is the only modern solution.

Ahrefs and 37signals show that owning infrastructure can be a smarter long-term investment for some. The cloud’s popularity does not make it the default best choice. Consider your project’s goals, budget, and tolerance for lock-in before deciding. Sometimes, the old-school approach of running your own servers can be the most forward-thinking move.

Sources

Category:

Year: