The hardest part of scaling a defense-hardware company isn't the hardware

Investors stopped paying for prototypes and started paying for throughput. Most defense-hardware teams have a shop floor and a group chat — not a coordination layer.

Eli Wood headshot

Eli Wood

July 15, 2026 3 min read
An abstract network layer of connected nodes floating above rows of factory workbenches, representing a coordination layer above a manufacturing floor

Problem

There has never been more capital chasing defense hardware. In the first five months of 2026, $14.6B flowed into the sector — more than all of 2025. But the money has moved: investors stopped paying for prototypes and started paying for throughput. Repeatable output. Supply chains that hold. A line that runs when the founder is asleep.

That shift catches a specific kind of company flat-footed: the one that can build a brilliant prototype and win a contract, but has never had to coordinate a hundred people, three suppliers, and a quality bar across a growing floor. The prototype was an engineering problem. Production is a coordination problem — and most of these teams have no coordination layer at all. They have a shop floor and a group chat.

Insight

The instinct is to buy an MES or an ERP and call it done. But the pain isn't on the machines — it's above them. It's the tacit knowledge trapped in one lead engineer's head. It's the supplier status nobody can see until it's late. It's the quality exception that gets caught only because the one person who knows happens to walk past. This is exactly the work that used to be done by hiring more experienced operators and program managers — people — and it's exactly the work a technology budget now has to absorb, because you cannot hire experienced defense-manufacturing PMs fast enough to keep up with a 10x ramp.

Applied AI helps here, but only if you apply it where it belongs. Four principles keep it from becoming another dashboard nobody trusts:

  • Keep a human in the loop where being wrong is costly. A missed quality exception on a defense system is not a place for autonomy. AI surfaces and flags; a person decides.
  • Separate the rules engine from the judgment engine. Compliance, sourcing rules, and process gates are rules — encode them hard. "Is this supplier about to slip?" is judgment — assist it, don't fake it.
  • Start where judgment is expensive and repetitive. The daily reconciliation of what-was-built vs. what-was-planned vs. what-shipped is the highest-value, lowest-glamour place to start.
  • Earn trust with explainability. If the coordination layer can't show its work, the floor will route around it in a week.

Path — what to do in two days

You don't need a twelve-month platform build to find out where this pays off. In two focused days you can: map the three coordination workflows that eat the most senior time; identify which are rules (encode) vs. judgment (assist); pick the single workflow where a thin AI layer gives the most hours back to the people you can't hire more of; and leave with a scoped, buildable plan and a rough ROI. Not a strategy deck — a decision about where to start.

The companies that win the throughput era won't be the ones with the best prototype. They'll be the ones that built the coordination layer above the floor before the ramp broke it.

If this is your situation, spend two days with us. We call it a Foundation Sprint.

About the author

Eli Wood headshot
Eli Wood

CEO, Black Flag Design

Eli Wood leads Black Flag Design, a creative technology company focused on shipping ambitious digital products, AI systems, and design-forward software with a direct point of view on how technology changes work.

Black Flag Journal

Field notes from building AI-native software in the real world.

Get the Journal

Related stories

More from the journal

Pen-and-ink sketch of a small clockwork robot working at a tool-covered workbench late at night while a human sleeps peacefully on a couch in the background, a wall clock reading 2:00 above
ai April 24, 2026 13 min read

The Agent Stays Up Late, Not Me

Every senior engineer knows the right way to set up a codebase. None of them do it. Here’s the four-stage framework we use — The Ratchet — to take a vibe-coded project all the way to a thing you’d trust in production, and the punchline about why this only just became worth doing.

Most teams have always known they should be running tests, type-checking, security audits, accessibility checks, dead-code analysis, prose linting, and a coverage floor. Most teams run two of those. Here’s why that math has finally inverted, and the four-stage framework we use to ratchet a vibe-coded project to a hardened one.

Keith Pattison, Founder of Black Flag Design

Keith Pattison

AI Builder @ Black Flag Design

Read
What a Year of Claude Code Trails Tells You About Your Team editorial image
claude code April 20, 2026 5 min read

What a Year of Claude Code Trails Tells You About Your Team

Claude Code leaves evidence — sessions, commits, PRs, review notes. Read it like a logbook and you'll find what devs actually need to know before they go deeper.

After a year of shipping with Claude Code across real client work, the signal isn't in any single session — it's in the trails. Here's what those trails told us about where Claude Code shines, where it drifts, and the habits devs should build before they lean in harder.

Eli Wood headshot

Eli Wood

CEO, Black Flag Design

Read
The Black Flag Playbook: Six Principles for Shipping with AI editorial image
playbook April 20, 2026 6 min read

The Black Flag Playbook: Six Principles for Shipping with AI

Battle-tested principles for teams building real software with AI-generated code. Human judgment, tight scope, and weekly evidence — the disciplines that keep AI-built systems reliable.

The six rules we use to ship production software with AI. Small scope, weekly demos, human-led oversight, and continuous improvement — drawn from six months of real client engagements.

Keith Pattison, Founder of Black Flag Design

Keith Pattison

AI Builder @ Black Flag Design

Read
Get the Journal