Why Diamond Foundry Treats Its Reactor as a Production Function

If Diamond Foundry considers reactor and process control central to its advantage, why has it not publicly built its strategy around selling MPCVD systems to other growers?

The answer may be that a one-time equipment sale is less valuable than keeping the reactor inside a compounding loop of process data and material learning.

Diamond Foundry says its equipment and process technologies were developed together and iterated in parallel. Public information does not prove that the company would never sell a reactor, but its visible path treats equipment primarily as part of internal production.

MPCVD equipment is not an isolated machine

Chamber architecture shapes the plasma. Microwave coupling changes energy distribution. The stage controls temperature gradients. Gas flow affects reaction boundaries and impurity transport. Cleaning and maintenance alter chamber condition. A material defect can result from several of these factors interacting.

A standardized equipment product must package enough know-how into specifications, controls, documentation, training, and service that a customer can reproduce the promised result. For high-end single-crystal diamond, many of the most valuable answers may still be inseparable from the supplier’s recipes, diagnostics, maintenance history, and operator judgement.

Internal use preserves a faster feedback cycle

When the same organization owns hardware and material output, a defect can trigger a chamber modification, sensor change, recipe adjustment, or maintenance experiment. The result returns as new material data. Over time, the system learns.

I use “production function” to describe that combined capability:

reactor architecture + process window + sensor data + maintenance system + operating knowledge + material feedback.

The barrier is not possession of a machine. It is the ability to improve predictable material output through the whole function.

Selling the reactor could create revenue, but it would also require freezing a supported configuration, exposing more process knowledge, and solving customers’ coupled equipment–process problems. For a company seeking value in wafers and application platforms, retaining the loop may offer greater long-term leverage.

Equipment suppliers remain essential

This does not make the external equipment model invalid. A specialist supplier can turn accumulated reactor knowledge into a reproducible platform and support many research or production customers. Deep collaboration between an equipment company and a material expert can also transfer application knowledge without one company owning every step.

The practical question for a grower is not “build or buy?” in isolation. It is whether the purchased platform can be understood, instrumented, modified, and connected to material specifications and customer evidence.

Treating a reactor as fixed capacity leads to equipment counts. Treating it as a production function leads to learning, control, and repeatable output.

Continue the series

• Why CVD diamond still needs vertical integration
• How MPCVD manufacturing maturity should be evaluated
• Explore Wonsen’s research areas