Making crystal growth predictable, one furnace at a time.

Crystal growth sits upstream of every chip and every optical device, and it is still run largely on experience and hindsight. We are building the tools that let growers see what their own data already knows.

Expertise from

  • Johns Hopkins University
  • Massachusetts Institute of Technology
  • UC Davis
  • UC San Diego
  • UC Santa Barbara

Our mission

A growth run takes weeks to months and gives up almost nothing until it is finished. That is why recipes are guarded, why a good hot zone is worth so much, and why scaling a process is slow.

We think the fastest way to change that is to start with what growers already have. Spotting the trends behind failed runs, watching a growth while it is still running, and catching sensor drift before it costs a boule are all possible today, on existing process data.

Each one earns its place on its own, and together they build toward a digital twin of the furnace that can be asked a question before a run is committed to it.