Defense tech scales on a new financial operating model, not just better technology.

Christian Brose's argument in The Kill Chain is that the United States optimized for small numbers of exquisite, expensive platforms while the variable that actually decides conflicts became the speed of the loop that senses, decides, and acts. The future force is large numbers of cheaper, attritable, autonomous, networked systems. That shift rewards software, sensor fusion, and open architecture over single heroic platforms, and it rewards organizations that can iterate at software speed.

Unit X describes how promising technology dies in the gap between commercial speed and Pentagon acquisition. The Planning, Programming, Budgeting, and Execution system moves on multi-year cycles with rigid color-of-money rules, while cost-plus contracting pays for effort rather than outcomes. New entrants rarely fail for lack of capability. They fail because the capital and contracting path cannot keep pace with the technology.

Anduril's core innovation is as financial as it is technical: build products with private R&D capital, sell finished systems commercially, and carry the development risk instead of passing it to the taxpayer. Lattice turns autonomy and sensor fusion into a software platform, and the company invests ahead of demand in mass manufacturing. It is the same pattern SpaceX runs on — first-principles cost, deep vertical integration, and a factory treated as the real product. Skunk Works ran a version of it too, with small autonomous teams that compressed the time from idea to fielded capability. It is no accident that Christian Brose, who diagnosed the problem, now sets strategy at the company built to answer it.

When a company bets its own balance sheet and still has to deliver government-grade assurance and speed, capital allocation, unit economics, program execution, and governance evidence become the real operating system of the business. Velocity without financial discipline burns the balance sheet; discipline without velocity loses the kill-chain race. The organizations that win build a control plane that holds both at once, connecting capital decisions, operating metrics, AI-assisted workflows, and audit-ready evidence in a single loop.

I build finance-grade AI operating systems for exactly this constraint: control-plane tooling that connects strategic finance, operating metrics, AI workflows, governance evidence, and executive decision loops. The goal is to let regulated, capital-intensive, high-velocity organizations move at software speed without losing auditability or capital discipline. It is the operational backbone the new defense model runs on.