One Big Machine

May 28th. Milan. I wasn’t sure what to expect.

I’d heard of the AWS Summit before — the kind of event that shows up in LinkedIn feeds and tech newsletters. Big names, big booths, the usual. I went in curious but calibrated. I left with something I didn’t quite anticipate: a new sense of scale.

I went with Davide, CTO at Enernext. Having someone who’s been deep in infrastructure for years next to you changes how you listen at these things — you process differently, ask different questions, notice different details.

The venue was enormous. Pavilions, dedicated speech areas, a whole floor of partner stands — Vercel, Cloudflare, Red Hat — companies that, when you think about it, owe a significant part of their architecture to AWS. They weren’t just sponsors. They were proof points.

We moved between sessions throughout the day. Chalk talks on DevOps. A deep-dive into Amazon Bedrock. CloudWatch internals. EC2 and ECS patterns. The anatomy of AI agents. Multi-agent orchestration. RAG pipelines.

It’s an AWS conference, technically. But AI was the current running through everything.

The thing that stayed with me wasn’t any single session.

It was a feeling that built slowly across the day: AWS is not a collection of services. It’s one big machine. Every piece fits into the next — EC2 feeds ECS, ECS connects to Bedrock, Bedrock slots into your agent, your agent queries a knowledge base backed by CloudWatch. The seams are intentional. The ecosystem is the product.

I’d understood this intellectually before. I felt it there.

The concrete thing I’m bringing home is an appetite for multi-agent systems with RAG — building pipelines where agents aren’t just autonomous but informed, grounded in real data, orchestrated toward something specific. There’s a version of that relevant to what we’re building at Enernext. I want to find it.

One day. One summit. A lot to think about.

Worth the trip.