Kilbride Systems
Find the bottlenecks that AI creates. Build the companies that solve them.
Kilbride Systems is a research lab identifying the physical, industrial and infrastructure constraints that tighten as AI scales, then building companies around solving them.
- Research lab
- Company building
- Early stage
- Founded 2026
Founder
Built from first principles.
Conn Martin
Founder, Kilbride Systems
Kilbride Systems was started by Conn Martin to work on a problem that sits underneath the entire AI build out: software is scaling far faster than the physical systems it runs on can be built.
Rather than starting with a product and looking for a use, the lab starts with the constraint. Energy, power infrastructure, cooling, water, land, materials and logistics come first, then the question of what company could actually solve one of them.
The bottlenecks
-
01
Energy
Generation, cost and availability of power at the scale AI is demanding.
-
02
Grid
Network capacity, connection queues and the infrastructure that moves power.
-
03
Capacity
Data centre supply, land and where new compute can physically go.
-
04
Thermal
Cooling systems, heat rejection and the water they consume.
-
05
Supply
Transformers, components, materials and the logistics that deliver them.
How the lab works
Research is the input. Companies are the output.
Every thesis moves through the same sequence. A bottleneck is only worth building around if it survives each step, and most will not. The point is not to publish research, it is to find problems that deserve a company.
Early stage, actively researching
The sequence
The thesis
The software scales.
The physical world doesn’t.
AI does not exist purely in software. Scaling it creates enormous demand for scarce physical systems: energy, grid capacity, data centre capacity, cooling, water, land, materials and industrial supply chains.
Every constraint that tightens becomes a problem someone has to solve. Kilbride Systems researches those constraints to find the ones worth building a company around.
Contact
Working on the physical side of AI?
I’d like to hear from researchers, engineers, operators, investors and founders working on energy, infrastructure and the industrial systems that AI depends on.