Applied AI — Manufacturing — Defense
AI-powered manufacturing systems to supercharge America's industrial base,
powered by foundational robotics models.
Foundational Vision-Language-Action (VLA) models interface directly with legacy 5-axis CNC machining cells. Inspect autonomous passes on an aerospace titanium seawater valve body below.
“Over the next decade, the Navy alone needs 250,000 new shipyard workers. The average defense machinist is 55 years old. The math is simple and brutal: American Dynamism is betting billions on a blue-collar workforce that does not exist.”
Silicon Valley knows how to take defense prototypes from Zero to One. But when thousands of VC-backed defense startups win contracts and need to field 50,000 units each to win a war, the strained American manufacturing base breaks under the pressure.
Submarine and surface vessel production at Newport News and Electric Boat face severe, compounding labor gaps.
Projected unfilled manufacturing positions across American industrial defense facilities through 2033.
Decades of irreplaceable machining know-how is retiring rapidly with near-zero backfill from Gen Z.
One trained operator orchestrates ten autonomous CNC and robotic workcells across unattended 24/7 dark shifts.
Engineered to retrofit into America's existing machine shops and prime contractor facilities with zero factory downtime.
A deployable robotic cell that retrofits onto any existing Haas, DMG MORI, Mazak, or Okuma 5-axis mill. Features pneumatic quick-change chucks, multi-axis robotic part loading, and automated chip evacuation.
A Vision-Language-Action model trained on physical machining physics, cutting dynamics, and sensory telemetry. Reads GD&T drawings directly to execute multi-axis toolpaths without brittle programming.
Optical and laser interferometer sensors inspect parts while the spindle is running. Measures sub-micron dimensional drift and adjusts toolpaths dynamically to prevent scrapped parts before they happen.
Centralized software managing shift scheduling, raw stock supply chains, dynamic tool wear replacement, and automated compliance manifests across fleets of autonomous machining cells.
From nuclear submarines to hypersonic glide vehicles, Nidus autonomous workcells produce the complex, high-tolerance components that define modern deterrence.
High-pressure titanium seawater valves, hull penetrations, and propulsion shafts for Virginia and Columbia-class submarines.
Scramjet fuel injectors, Mach 5+ thermal protection nosecones, and thin-wall solid rocket motor casings.
High-volume tactical UAV airframe spars, sensor gimbals, and kinetic interceptor fins capable of scaling to 50,000 units.
Precision star tracker brackets, optical benches, and low-thermal-expansion satellite bus structures for resilient LEO constellations.
Our team bridges the cultural divide between high-volume physical manufacturing and frontier AI. We have deployed national security systems, built factory line automations, and trained foundation robotics models.
Started as a manufacturing engineer at Robert Bosch on America's factory assembly lines. Former deployment strategist and operations lead at Palantir, Peregrine, and Vannevar Technologies. Author of the defense industrial manifesto Why American Dynamism Will Fail.
Directs autonomous robotics architecture and foundational Vision-Language-Action (VLA) model development. Background in frontier AI, real-time spatial intelligence, and robotic manipulation systems engineered to perform outside controlled lab environments.
Specializes in sub-micron closed-loop optical metrology, laser interferometry toolpath compensation, and real-time inference on edge tensor hardware. Led physical autonomy projects across defense primes and academic research centers.
Expert in 5-axis DMG MORI and Haas CNC machine retrofits, pneumatic fixture chuck integration, and high-temp alloy machining (Inconel 718, HY-100 steel, titanium). Directed precision manufacturing lines for aerospace rocket propulsion.
We work together in-person in NYC. Great team, fast iteration, daily team lunch. Autonomous manipulation in real manufacturing environments is an unsolved engineering problem — and the most important one in America.
We prefer physical ground-truth over synthetic benchmarks, and shipping over theorizing. We keep our technical staff small and senior.