Operations | Monitoring | ITSM | DevOps | Cloud

AI agents can write code and run tools on their own now. The problem is they still need somewhere to actually do it. exe.dev's David Crawshaw joins Michael Reid to break down why persistent virtual machines might become the backbone of agentic AI, and what happens to cloud economics when one person is running dozens of agents at once.

What's currently held together by duct tape and prayer in your stack? Drop it in the comments and let's fix it.

You migrated the workload. The app's live across clouds. But is the traffic between them actually locked down, or just assumed to be? There's a layer of the network doing the heavy lifting here, and it goes by a name that gets confused with something else constantly. Full breakdown on our blog, link in bio.

With Megaport, you get global networking, GPU-powered bare metal, and now object storage powered by Wasabi Technologies. What other shifts are you seeing in AI infrastructure right now?

What happens when AI agents stop simply answering questions and start using computers of their own? It could create an entirely new boom in virtual machines. In this episode of Uplink, David Crawshaw, Co-Founder and CEO of exe.dev, joins host Michael Reid to explore the infrastructure behind the rapidly emerging world of AI agents. As agents become capable of writing code, running applications, operating tools, maintaining state, and working independently, they need more than access to an AI model. They need computing environments where they can actually get work done.

Our Megaport technical expert, Steve Tu explains what IPsec is, how it secures network traffic, and where it’s used across VPN and cloud connectivity. About Megaport Deploying infrastructure should be fast and simple. Megaport’s software-defined platform provides private compute, network, and storage — so you can build secure, scalable infrastructure for cloud, enterprise, and global AI inference workloads. Trusted by the world’s leading enterprises, Megaport operates across 1,200+ enabled locations globally.

As infrastructure gets more complex, with workloads spread across clouds, regions, and providers, every hop your data takes between them adds up. Where do these costs actually come from? Has your team ever traced a surprise bill back to data movement? With Megaport, cloud egress costs are minimized by routing data privately between providers, avoiding the higher transfer rates associated with public internet routing.

AWS, Azure, Google Cloud all excel at different things, and the right pick depends on where your company already lives. Whichever you pick, the network is what ties it together. That's exactly what Megaport Cloud Router does.

Someone else in the deal can. Megaport gives DC operators, MSPs, network providers, and system integrators one network layer to connect AWS, Azure, Google, and Oracle in minutes.

Ben Baldieri, founder of The GPU, joins Megaport CEO Michael Reid to break down the money, the chips, and the deals behind the AI infrastructure boom. From 5-gigawatt megasites, to the smaller assets Europe is using to compete. The full episode is now live on our channel.