A Future for Abundant Manufacturing
Most of the stuff around you—materials, medicines, coatings, electronics, fuels—comes from giant, expensive, centralized factories. They’re slow to build, slow to adapt, and locked into specific products.
We’re working toward something very different: a world where complex manufacturing is ultra‑cheap to run, cheap to build, and can be deployed almost anywhere in modular, standardized systems.
Here’s how we do it.
The Big Idea: Proteins as Tiny Machines
At the heart of this plan is a simple thesis:
The leverage point in manufacturing is at the nanoscopic scale—where molecules are built and broken apart. And the most practical tools at that scale are proteins.
Proteins are nature’s tiny machines. They’re how your body builds new molecules, breaks down food, repairs DNA, and more. Each protein’s function is determined by its sequence (the order of amino acids) and its structure (the 3D shape that sequence folds into).
If we can reliably map sequence + structure → function, we can start to design proteins that:
Build new kinds of materials
Catalyze chemical reactions cleanly and efficiently
Bind to very specific molecules (great for purification, sensing, or targeting)
And if we can do that at scale, we can rewire how manufacturing works.
Step One: The Protein Function Engine
The first part of the strategy is to build what we think of as a Protein Function Engine.
You can think of it in three layers:
Screening at scale
In the lab, we test very large libraries of proteins against many possible functions:Can this protein speed up this reaction?
Can it bind to this molecule?
Can it tolerate heat, solvents, or pressure?
The goal is to generate a huge, rich dataset of “this design → this function (or set of functions).”
Mapping design to function
From this data, we build an index that connects:
“Here’s the protein design” → “Here are all the useful things it can do.”
One protein might catalyze more than one reaction, or bind multiple targets. Capturing that complexity makes the system more powerful over time.Training models that can design new proteins on demand
Using all that data, we train models that take as input:
“We want a protein that does this function under these conditions.”
…and output a candidate protein sequence likely to perform that job.
We started with natural amino acids (the standard building blocks used in biology). Over the long run, we’ve begun to extend into non‑canonical amino acids—synthetic building blocks that expand what proteins can do, like adding new chemistries or higher durability. Our protein function engine now has hundreds of thousands of reactions we’ve studied with thousands being added every week. We’ll continue to scale this to millions of reactions every iteration.
Step Two: From Lab Hits to Kilograms—Fast
Having a clever protein isn’t enough. To matter in the real world, those proteins have to be:
Stable in real process conditions
Easy to plug into repeatable hardware
Scalable from nanograms in a test tube to kilograms in a reactor, and metric tons at industrial scale
That’s where our Catalytic Reactor Program comes in.
Industrializing proteins
We’re building general methods to:
Immobilize and stabilize proteins so they keep working under industrial conditions (temperature, pressure, solvents, etc.).
Drop different proteins—enzymes, binders, and more—into standardized reactor setups rather than inventing a new process every time.
The idea is to make the reactor platform function‑agnostic. In other words, the hardware and operations shouldn’t care which protein is inside, as long as we follow some shared rules.
Rapid scale‑up
Our target is aggressive:
Take a new protein function from lab discovery to kilogram‑scale in weeks, not years.
That speed matters because it lets us:
Test new materials and products quickly
Iterate on process conditions
Generate more performance and economic data
Feed better information back into the models
This is how the lab → factory gap starts to close.
Step Three: Revenue Flywheel & Market Expansion
To sustain this kind of deep R&D, we need to build products on top of it. So we’re starting where our tech has immediate, painful problems to solve.
Beachhead: Better Materials via Prepolymers
Our initial focus is on materials built using prepolymers
Very roughly:
Polymers are long chains of repeating molecular units (think plastics, resins, fibers).
Prepolymers are molecules that are just about ready to polymerize, but only do so based on a specific condition (temperature, eventually impact/radiation)
By using our proteins to build these structures, we can aim for:
Faster printing
Higher mechanical strength
Better thermal performance
That’s especially valuable in defense and other demanding applications, where performance and reliability are absolutely critical.
And with our first customers buying material that will scale to ten tons per month within a year, we’re off to the races. That’s the revenue engine that funds the deeper tech.
The Endgame: “Dreadnoughts”
All of this is building toward a larger goal: what we call Dreadnoughts.
A Dreadnought is a modular, shipping‑container‑scale manufacturing unit that:
Integrates Aether’s protein catalysts
Uses our standardized reactor systems
Comes as a reconfigurable chemical manufacturing platform
Instead of a single massive plant specialized for one product, imagine:
Deployable, containerized units that can be re‑tooled via software and “swapped‑in” proteins
Lower OpEx (operating expenses) thanks to biocatalysis and automation
Lower CapEx (capital expenses) because we control and standardize the plant architecture itself
An example path: minerals
One illustrative path is mineral‑related processes:
Start with very high‑value targets like germanium, platinum, gold.
As costs fall, extend to lithium, titanium.
Eventually reach copper‑scale markets—massive, multi‑hundred‑billion‑dollar spaces.
The key is that the same underlying platform and design engine power all of it.
What Has to Be True
For this vision to work, three things must be true:
Universal protein function model
We need models that can reliably design proteins for a wide range of molecular functions regardless of whether they exist in nature already.Fast, generalizable reactor platform
We need containerized reactors that can take new functions from lab conditions to kg‑scale to metric ton scale quickly and repeatably.Product driven revenue engine
We need products and markets that can grow the core business to:Fund ongoing R&D
Reinforce our data, models, and scale advantages
Keep us several steps ahead technologically
The Path ahead
Aether has already made extraordinary progress on all three fronts, we’ve generated dozens to hundreds of new chemistries, we’re scaling our first product line to 10s of metric tons a month, and we’re selling our first product as fast as we can make it.
As we continue to grow, we’ll invest more in our core protein function models, expand their capabilities, build more and more generalizable catalytic reactors, all in preparation to build a future of abundance for the human race.