Everyone Can Build Software Now. That's Why Cloud Matters More

There is a story going around that AI code generation is about to make cloud infrastructure less relevant. The reasoning sounds clean. If anyone can build software by describing it in plain language, then the people who deploy and run that software matter less, and the cloud they run it on fades into the background.

As someone who works with AWS most days*, I suspect the opposite is closer to the truth. The AI coding boom is quietly turning into a cloud boom. The barrier AI is lowering is the barrier to writing code. It does very little to the barrier that comes right after, which is getting that code running where other people can actually use it, without it falling over. If anything, the flood of new software makes that second barrier matter more, not less.

What is actually happening

Tools like Cursor, GitHub Copilot, or Claude Code let people produce working software from ordinary sentences. Vibe coding platforms take it further, so someone can describe the app they want and watch it appear. The person doing this no longer needs years of training to get a first version running.

The interesting part is who is now building. It is not only professional engineers going faster. It is product managers building the internal tool they always wished existed. It is marketers spinning up a landing page and a small automation. It is analysts turning a messy spreadsheet into a proper little application. People who once had to wait in line for engineering time can now build for themselves.

What excites me about this is who it helps. The amount of software being created is growing fast, and a lot of it now comes from people who never had a way to build before. That is a good thing. When the person who understands a problem can also shape the solution, good ideas stop waiting in a queue for scarce engineering time.

Building is not shipping

Building code is only half the job. Code has to run somewhere before anyone else can use it. Some AI tools will even deploy a first version for you now. What they do not do is make it ready for the wider world: secure, able to handle load, sensible on cost, and connected to your own data. That is the part that decides whether you built a demo or a product.

Think about what a working application actually needs. It has to be reachable from a browser or a phone, on devices you do not own and cannot touch. It has to stay up when you close your laptop. It has to serve more than one person at a time, keep their data safe, and still be there tomorrow. A personal machine was never meant for that. Serving a team, an organisation, or customers takes somewhere built to stay online and grow as more people use it. The AI wrote the recipe. Someone still has to run the kitchen. That someone is the cloud.

Why cloud gets more important, not less

More builders means more applications. More applications means more things that need to be hosted, scaled, secured, and kept alive. And most of these new builders would rather spend their time on the problem they are solving than on running the infrastructure underneath it, so they want systems that simply work.

This is where cloud services like AWS earn their keep. A person who just used AI to build an app does not want to configure a server, patch an operating system, or run their own database. They want to point at a service that handles it. Managed databases, storage, authentication, and on demand compute become more valuable in this world, not less, because the audience for them just grew from a small technical audience to almost everyone who builds.

The story says AI removes the need for the people in the middle. What I keep seeing is that the demand did not disappear. It moved one step down the stack, to where the code meets the cloud.

The skills I think matter more now

If that is the direction, then the skills that matter shift too. These are the ones I keep coming back to:

1. Knowing the managed building blocks. Understanding what the common managed services for compute, databases, storage, and logins are for, and reaching for a managed option instead of building the plumbing by hand.

2. Deploying carefully. Getting an application from a working idea to a live, reachable service without cutting the corners that lead to outages or leaks.

3. Keeping an eye on cost. Many small applications add up, so it helps to design with the bill in mind as the number of little apps grows.

4. Security fundamentals. Least privilege, protecting secrets, and safe defaults. With far more software being shipped, more of these gaps appear simply because there is more of everything, and they are worth helping to close.

5. Observability and monitoring. Being able to tell when something breaks, and see why, before the people using it are affected.

6. Making the cloud approachable. Explaining all of the above in plain terms, so anyone building can do it with more confidence.

None of these are about writing more code faster. They are about judgment, safety, and running things well. That is the half of the job AI is not taking.

Where I see this most clearly: international development

I see this most clearly in the world I work in, international development. The push to adopt AI, automation, and digital systems is accelerating fast. But the value never comes from generating the code or spinning up a model. It comes from doing it responsibly and then actually running it. Automated business workflows, well organised institutional knowledge, a management information system that connects activities to the data behind them, and capability uplift so program teams can build with confidence.

And because these are development organisations accountable to the communities they serve and the partners who fund them, the running half carries extra weight. Governance, security, inclusion, safeguarding, and value for money cannot be an afterthought bolted on at the end.

Building is the easy part now. Making it land, safely and at scale, for teams whose main job is not running infrastructure, is the harder half of the job, and it is exactly where cloud and digital professionals become essential.

The opportunity

Not every vibe coder needs to become a cloud expert. But every organisation that starts deploying AI generated tools suddenly needs cloud thinking, whether it planned to or not. The tools multiply, the hosting questions multiply with them, and someone has to answer them well.

The professional who understands both sides, how AI generated code gets made and how real cloud infrastructure runs, becomes the bridge between the two. Cloud skills are not being replaced by this wave. They are being asked to serve a much wider audience than before. AI democratised creation. It did not democratise distribution. Everyone can build software now. Someone still has to host it, secure it, watch it, and keep it running. That someone has never been more needed.

Syaeful Bahri is an AWS practitioner and subject matter expert.