Silicon Valley has a new favorite job title: Forward Deployed Engineer.

Microsoft just committed $2.5B to build out a forward-deployed engineering organization. AWS announced a $1B investment of its own. OpenAI and Anthropic are also in the game. If you’ve been following the Tea, you know I’ve been brewing on this trend:

So what exactly is a Forward Deployed Engineer?

Think engineer, consultant, and product manager rolled into one. Instead of building software from headquarters and handing it off, they embed directly with customers—learning how the business actually works, optimizing workflows, and making the technology useful in practice.

Palantir arguably pioneered this model years ago. They knew their technology was complex, so they sent engineers into the field to make it work. The model paid off: deeper integrations, higher switching costs, and incredibly sticky customers.

It's becoming increasingly clear that implementation—not frontier models—is the next strategic frontier in AI.

You can already see it in the OpenAI–Anthropic battle. A frontier model is only as valuable as where—and how deeply—it gets deployed. And as these companies eye the public markets, recurring revenue and customer stickiness matter just as much as model performance. Honestly, maybe more.

Let the FDE games begin ☕️