Why we started

Most companies pay more for their infrastructure than for their product.

Every company running software pays for two things: the thing they sell, and the machinery that keeps it alive: servers, orchestration, pipelines, people on call. It has been that way so long that nobody itemises it anymore.

None of it was designed. It accumulated. Each layer solved a real problem and left a dependency behind, and the next was built on top of it. Today seven to nine layers of software sit between your code and the machine that runs it. None of them are your product.

We started LILY because that stack can't be fixed from the inside. Optimising a layer still leaves the layer.

What we built

We compile programs, not images.

A normal build flattens your code into a binary and the structure disappears. Ours keeps it: functions, methods, the boundaries already in your code stay visible after compilation. The runtime knows which parts hold state and which don't.

Everything follows from that. Parts that hold no state run in parallel, move between machines, or swap into a live system. Parts that do are snapshotted and restored exactly as they were. Nothing needs a container around it. The units are already separate.

We run directly above bare metal, at the level of executable instructions. No hypervisor, no guest OS, no orchestrator.

Join us on

The Founders

Meet the founders.

LILY started as infrastructure for research workloads. Most of the work went into the layers underneath rather than the actual computation. Provisioning, orchestration, capacity idle between runs.

Those layers weren't load-bearing. We removed them and built the rest ourselves. That became the product.

Leon Kaiser

Milan Wiedemann

Investor

Backed by Mätch VC.

A deep tech fund out of Stuttgart, carried by around ninety German industrial family businesses. They know the industry our work has to hold up in, which means we rarely have to explain the hard parts. We are grateful to have them with us.