Moment 01 / 05

AI tools + low-cost hardware are unusually powerful.

A small team of non-experts can do anything.

Every company is on the disruption chopping block.

All incumbents have legacy baggage and inertia. The playing field is unusually open.

The compounding effects are unusually high.

Platforms that bridge software, AI, and physical operations lock in fast - every node, every customer, every interaction deepens the moat.

Atoms 02 / 05

The next trillion-dollar companies will control reality

Code is becoming infinite. Intelligence is becoming abundant. Scarcity is shifting to physical execution. The next great platforms will orchestrate factories, logistics, energy, robotics, and critical infrastructure - compounding through deployed learning systems that build memory and get smarter with every cycle.

Window 03 / 05

A first-mover window is open for the next generation of platforms

Software platforms defined the last era. The infrastructure layer connecting AI to physical operations is unclaimed. The next trillion dollar companies will combine AI, hardware, and real-world systems into global platforms that grow stickier with every deployment.

Flywheel 04 / 05

These platforms will compound like the great digital platforms - but with deeper moats

Like the best pure software platforms, they scale through distribution, network effects, and data. But they also accumulate proprietary operational memory that continuously improves the AI systems orchestrating the platform with every deployment.

Where 05 / 05

The Midwest is structurally underpriced for what comes next

AI has collapsed the premium on elite coastal engineering talent. When world-class technical output no longer requires world-class payroll, advantage shifts to regions with lower costs, industrial infrastructure, and proximity to physical operations.

We back platforms, not projects.

What fits our thesis

  • Repeatable hardware that can be deployed at low cost across many sites.
  • A unifying software layer that captures long-term value.
  • Workflow integration that creates real, operational switching costs.
  • Network effects from usage, data, and standardization.
  • Centralized intelligence - AI models, memory, and operating context.
  • A customer base broad enough for global proliferation.
  • Hybrid economics: software-like margins, infrastructure-grade defensibility.

What doesn't fit our thesis

  • Businesses dependent on a handful of giant buyers.
  • Impressive but non-repeatable, one-off integrations.
  • Highly custom integrators with weak productization.
  • Hardware with no compounding software layer.
  • Tools most likely to be acquired before reaching real scale.
  • Pure-cloud stacks with no physical embedment or real-world feedback.