The cost-per-kill calculation is becoming the most important number in defense procurement — and two contracts awarded this week make that case better than any policy document could.
Perennial Autonomy's $500 million JIATF-401 contract is the headline, but the number that actually matters is buried in Army Secretary Driscoll's congressional testimony: the Merops interceptor costs roughly $15,000 per unit, while the Iranian Shahed it's designed to kill runs between $30,000 and $50,000. That's not a capability story — it's an economics story. The Pentagon is finally buying attrition math that works in America's favor.
Perennial's trajectory is worth pausing on. Originally launched by former Google CEO Eric Schmidt as Project Eagle, it developed the Merops for Ukrainian forces countering Russian Shaheds. The U.S. military is now deploying the same interceptors against Iran's variants. That's a real-world operational feedback loop compressed into roughly two years — the kind of iteration cycle that legacy primes simply cannot match.
The Shield AI selection adds a different dimension. The Pentagon tapped Shield AI to integrate its Hivemind autonomy software onto LUCAS — the one-way attack drone reverse-engineered from an Iranian Shahed 136 — with a swarm demonstration planned for this fall. A single operator retaining strike authority while Hivemind coordinates autonomous maneuvering across multiple platforms isn't science fiction anymore. It's a pilot program with a demo date.
What connects these two awards is the same underlying logic: the Pentagon is no longer buying platforms. It's buying cost structures and software architectures. LUCAS carries a roughly $35,000 price tag and can be swarmed by a single operator. Merops intercepts threats that cost more to build than the interceptor itself. Both bets only make sense if you're planning for volume — which is exactly what the FY27 DAWG budget request of $54.6 billion signals.
Meanwhile, Mach Industries' $50 million acquisition of solid rocket motor startup Exquadrum is the supply chain move that makes all of this possible at scale. The domestic SRM market is effectively controlled by two large primes, with lead times stretching years. Mach isn't just solving its own bottleneck — it's positioning Mach Energetics as infrastructure for the broader defense tech ecosystem. That's the kind of vertical integration bet that looks obvious in retrospect and contrarian at the time.
The pattern across all three developments is the same: the Pentagon is rewarding companies that solve the production problem, not just the capability problem. A drone that works brilliantly but can't be manufactured at volume is a prototype. The contracts going out right now are for the companies that figured out the difference.