Blackbody
Diamond thermal hardware for compute in orbit. Moving kilowatts off GPUs in a vacuum, where nothing else can.
We convert the world's idle gem diamond reactors into industrial capacity.
Diamond thermal hardware for compute in orbit. Moving kilowatts off GPUs in a vacuum, where nothing else can.
Diamond electrodes that destroy forever chemicals. Sold per gallon to the industries the EPA just put on the clock.
No new fab. The chamber already exists; we change what goes into it and what reads it.
A gas and power retrofit on an existing CVD chamber, plus an optical port. The operator keeps the asset.
Optical emission and substrate temperature go to a growth model that corrects the recipe mid run, before the crystal is spoiled.
The grade travels with the part and returns to the model as a labelled example. That loop is the compounding asset.
Carries heat, holds off voltage, survives chemistry that eats every other electrode. Nothing else does all three.
Log scale, decades marked. The multiplier is against the strongest competing material in that row.
| Property | Diamond | 4H-SiC | GaN | Si |
|---|---|---|---|---|
| Thermal conductivity W/m·K | 2200 | 370 | 130 | 150 |
| Bandgap eV | 5.47 | 3.26 | 3.40 | 1.12 |
| Breakdown field MV/cm | 10 | 2.5 | 3.3 | 0.3 |
| Electron mobility cm²/V·s | 2000 | 900 | 1250 | 1450 |
| Hardness GPa | 90 | 25 | 12 | 11 |
| Potential window V, aqueous | 3.0 | n/a | n/a | n/a |
Room temperature figures for high quality material.2 The potential window is what breaks carbon fluorine bonds instead of corroding the electrode.3
In vacuum there is no convection. A machine in orbit can only radiate, and radiator area is among the heaviest things it carries. Diamond moves kilowatts off the die so less area has to be launched.

Boron doped diamond has the widest aqueous potential window of any common electrode. That is what lets it break carbon fluorine bonds instead of corroding, destroying PFAS rather than moving it to another filter.

The highest breakdown field of any semiconductor, settled physics for decades. What blocks it is doped substrate quality, which is a materials problem, and that is the problem we work on.

Windows and spreaders that stay flat under beams that deform everything else, because heat leaves faster than it can distort the optic.

Reliable work starts with reproducible material. Characterised plates with defined orientation, purity and strain. The scan travels with the plate.

They already exist. Converting a chamber beats pouring a new one, and the operator keeps the asset.
Corrects the recipe mid growth. Normally you set the parameters, wait days, and find out afterwards.
Every graded plate is a labelled example. No single reactor owner makes enough volume to build that corpus. A fleet does.
Thermal parts tolerate defects electronics cannot. It is the grade the fleet hits today, and it funds the quality electronics need later.
No. We sell watts moved and gallons treated. The stone is the means.
Watts of heat, gallons per day, or a substrate spec.
We reply with a specification sheet. Privacy.