3 reasons AI needs platform engineering

Hey there! Welcome to Platform Weekly. Your weekly scuba through the platform engineering reef. Every week, we dive into best practices, news, lessons and updates from the platform engineering community.

But first… this Platform Weekly is partnering with Jellyfish, the Intelligence Platform for AI-Integrated Engineering:

Platform teams are now the architects of AI transformation. That means getting adoption right, enabling developers, and proving it's working.

Jellyfish analyzed 20 million pull requests across 200K developers at 700+ companies to find out what actually drives results. Top AI adopters are seeing 2x the PR throughput. But most teams are leaving a lot on the table.

• Explore their AI Engineering Trends report and benchmark your org against real-world data.

• Not sure how to build the foundation that makes AI scale safely and consistently? They have a guide that covers exactly that.

• Need to show leadership it's working? Request an AI Impact demo from them.

3 reasons AI needs platform engineering

If you’ve read anything from me or the wider community in the last 6 months you’ve probably heard “Platform engineering is fundamental to the success of AI” best exemplified by the famous Google Cloud stat where “86% of leaders believe that platform engineering is essential to realizing the full business value of AI”. 

This idea has been on my mind for awhile now, but spread like wildfire across the industry with the publishing of the State of AI in Platform Engineering Vol.1 and the 2026 DORA report.

What exactly does it mean when we say that? Let me break down 3 key reasons for it.

Platform engineering is the key (accidentally)

You don’t need me to tell you the benefits of AI. The ability to enable engineers to massively supercharge the creation of software, or to autonomously create software itself. It’s only possible to do this however if your security, your governance, your CI/CD, your testing, and all the other aspects of your software engineering organization are actually able to keep up with it.

It can only work if your team is locked in on making sure productivity is being properly measured, and if DevEx experience is actually being taken care of.

And guess what? These are all key elements of platform engineering.

There are 12 table stakes that define all successful platform engineering initiatives. If you talk to a mature platform team - there is a 99.99% chance that they cover these 12 points.

These are things like platform as a product, infra automation, testing, API management, observability, cost management etc. ALL platforms should have these.

And so it’s no surprise that orgs with platforms are the ones who are able to deliver the incredible value that AI promises - as AI needs all of these things to succeed.

From multicloud to… multiAI (infra)

AI systems depend on extremely complex, rapidly evolving infrastructure stacks including models, data pipelines, GPU compute, storage, and cloud environments. 

In the same principle as how your platform might enable multi-cloud, platform teams can build their AI infra as modular, building blocks covering compute, storage, networking, GPUs, and AI services. Devs are exposed through an abstraction layer in the IDP where they teams swap or integrate components without redesigning the entire stack. Your platform could enable teams to:

• Switch AI providers or models as capabilities or needs change

• Use different GPU resources or orchestrators depending on cost or performance

• Integrate new data sources or vector databases without rebuilding pipelines

• Run workloads across hybrid and multi-cloud environments

You can see immediately how this could drive an orgs capability. What if a new model from Anthropic is 10x better, cheaper or faster than an OpenAI or Google one? Just swap it. Does a new set of GPU resources cost 10% less? Then swap it. This kind of immediate hands-on freedom is a massive massive unlock.

Everyone becomes a developer

At a round table 10 months ago, a Head of Platform at a major US retailer told me “We are on the path to any office worker in the company spinning up their own specialized micro-apps through the platform”. The words of the conversation sounded basically sci fi, but as we talked through the platform engineering principles that would enable it - I knew exactly how it could be done. And how platform engineering was perfect for it.

Now… almost every second conversation I’m having is sharing that this is the path they’re on.

We have entered the age of the Enterprise Citizen Developer, where non-developer colleagues can use AI to vibe code their own specialised microapplications. One conversation 2 weeks ago, “Our CEO himself has coded out 17 micro-apps through the platform. Stuff he has always wanted to use for his own work, but didn’t think it was worth dev time”. That story is everywhere.

How does platform engineering enable this? We’ll go back to point 1 in this newsletter.

What does this idea need? It needs governance, security, policy, cost management, testing, change validation etc etc etc

These are all things that platform engineering provides.

So what does this mean for you? Well, if you’re a platform engineer… you’re a lot more important than you think. And if you're trying to get funding, or interest in your platform initiative, your boss might not be seeing the full picture.

So, make sure to send this to them;)

Quick bites

Event highlight:

• How to benchmark your platform: Quantify your pain (& solve it)

From the community:

• Want to watch the recordings of PlatformCon workshops and talks you weren’t able to attend? We’re uploading more every day on the community Youtube channel!