FDE Is Booming. But Is It Scalable

Over the past few months, Forward Deployed Engineers (FDEs) have become one of the hottest roles in AI. Palantir, OpenAI, Anthropic, Cursor — everyone's hiring. The feeds are full of posts on what they do, what they earn, and how to become one.

But FDE isn't new. Palantir scaled this model over a decade ago, and before the buzzword there were solution consultants and implementation engineers doing the same thing: embedding in customer environments, decoding business problems, and making technology actually work.

So the better question is this: if it's so valuable, why hasn't it become how software companies scale?

The Real Edge of FDE Work

The answer hides inside its greatest strength.

People assume a great FDE walks in, instantly understands the business, and produces the perfect solution. They don't. The real value is compressing the feedback loop. Ship something imperfect. Sit next to the customer. Watch it break. Fix it. Repeat — in hours, not weeks.

The edge isn't speed, and it isn't precision. It's proximity. That's why AI projects actually land: the hard part is rarely the model. It's the messy data, the broken workflows, the habits and details you only see up close. When a customer says "we want to improve efficiency," the most valuable person in the room is the one who knows what that actually means.

The Scaling Problem

So far, good news. Now the hard part: can it scale?

Every FDE win takes elite talent, deep customer involvement, heavy customization, and constant on-site iteration. That's not software — that's consulting.

Software's magic is simple: build once, sell a thousand times, marginal cost near zero. FDEs serve one customer, then the next, then the next. It's the question Palantir has faced for years, now landing on every AI company: are you a software company, or a consulting company wrapped in software? Double your customers, and do you double your FDEs?

That's not necessarily bad. But it's a different business.

Reuse Is the Only Answer

So the winners won't be whoever has the most FDEs. They'll be whoever turns what an FDE learns at Customer A — prompts, connectors, workflows, playbooks — into reusable assets for Customer B.

In the end, it comes down to one number: reuse.

Reuse it, and you're selling software. Start over every time, and you're selling people. And no matter how much AI is in the pitch, you're no longer scaling software.

You're scaling headcount.