Fermion
Autonomous fermentation control — temperature, pitch and feed that hold the yeast on profile.
Distilleries & Spirits
Caskent
The autonomous operations layer for distilleries and spirits production.
Overview
What Caskent does.
Caskent closes the perception-to-control loop on a distillery: autonomous fermentation, distillation-cut control, AI spirit-quality sensing, maturation optimisation and robotic bottling.
The product stack runs the batch from grain through cask. Each deployment produces labelled production traces that train the next control model, so the system gets safer and more precise with every run.
| Detail | Value |
|---|---|
| Stage | Seed |
| Category | Physical AI |
| Year | 2026 |
| Website | caskent.com |
| Headquarters | Not disclosed |
| Raised | $1.5M |
| Investor | 10xC |
| Model | Outcome-hybrid pricing per batch, entered through a paid pilot on one line. |
| Measured by | Batch consistency, spirit yield, energy per litre, lines expanded per customer. |
Products
One loop, module by module.
Every deployment on caskent.com writes labelled traces back into the model, so the system generalises across plants, alloys, grades and geometries.
Cutrix
Distillation cut control that reads the run and stops the still at the right point.
Nosera
AI spirit-quality sensing from still to bottling line.
Caskora
Cask maturation twin that predicts how a spirit develops in wood.
Bottlex
Robotic bottling and handling on the finished line.
Why we invested
The thesis behind Caskent.
Spirits are made by people who know the batch and the still. Scale means more brands, more warehouses and fewer of the people who can taste the difference. Caskent encodes the judgement in software that runs every batch to the same standard.
The buyer is a distiller paying in hard currency. The wedge is one line — fermentation or distillation cut — entered through a paid pilot, with the data flywheel funding each next module.
Who is building it
Laiba Nisar
Co-Founder & CEOLaiba owns the distillery relationships, the pilot programme and the commercial model.
Sharon Reddy
Co-Founder & CTOSharon owns the process models, the edge runtime and the bounded write-back into still and line controls.
Building something similar?
Building in Physical AI? We want to hear about it.