Thesis 02 · Materials & chemistry

Inverse Design for Matter

AlphaFold showed that structure can be predicted. The next step is designing matter that doesn't exist yet, and making it.

The protein playbook is to learn structure from data, generate candidates, and test only the most promising. It's now moving to inorganic materials and small molecules. Generative models propose stable crystal structures and molecules faster than any lab can evaluate them by hand.

The constraint has moved from imagining candidates to testing them. Whoever closes that gap opens a market no one can size yet.

01What happens when the lab runs itself?

Self-driving labs close the gap. A model predicts, robots synthesize, instruments characterize, and the results retrain the model. Each turn of the loop makes the next one smarter.

How a new material gets found
PredictSynthesizeCharacterizeRetrainWeeksper turn

Each turn takes weeks, and every result makes the next prediction better.

Discovery cycles that took years compress into weeks, and the field shifts from finding materials to specifying them: describe the properties you need, then design backward.

02What does this look like in production?

MatScience AI is a Harvard–MIT team that combines AI, researchers in the loop, laboratory validation and manufacturing to develop product-ready materials. Its flagship product, already in production, is a next-generation liquid coolant for high-density AI data centers.

According to the company, it targets savings of roughly $2–4M per gigawatt per year through better heat transfer, lower pumping power, less maintenance and a longer service life. The team is also developing advanced magnesium alloys for semiconductor and aerospace uses.

The point isn't the coolant. It's that a model-designed material reached a buyer with urgent demand.

03Where does it break?

A predicted material is worth nothing until someone makes it at scale and a buyer qualifies it, and that takes years. Most companies in this space will stall between the lab and the customer.

The winners will own the commercial path: who buys it, at what price, and how long qualification takes.

04How big is the market?

Like the biomarker thesis, this one has no TAM yet. We're entering a new phase of chemical and materials discovery where the use cases are still unknown. You can't size a market for materials that haven't been invented.

What I look for

  • The loop: do they own design, validation and manufacturing, or hold a clear wedge into it?
  • First product: is it aimed at a buyer with urgent demand, like data-center cooling, batteries or semiconductors?
  • Data: is the experimental data proprietary, and does every loop turn add to it?
  • Path to scale: who makes it at volume, and how long does the buyer take to qualify it?

What would make me wrong?

  • If platforms that license software to chemical majors win instead of companies that make materials themselves, the slow path to scale matters less than I think.

Diligence questions

What I ask founders building against this thesis. Open any question to see why it matters.

ILoop

How fast does discovery actually turn?

  1. 1.Which steps of predict, synthesize, characterize and retrain do you own?

    Owning the full loop is the moat; renting half of it is a services business.

  2. 2.How many loop turns do you run per month today, and what limits it?

    Cycle time is the metric the whole thesis rests on.

  3. 3.What in your experimental data can't be reproduced from public databases?

    Models commoditize. Proprietary measurements don't.

IIProduct

Is there a buyer who needs this now?

  1. 4.What's the first material, and which spec does it beat?

    A product is a number a buyer already tracks, improved.

  2. 5.Who's the buyer with urgent demand, and what does qualification involve?

    Qualification time is the hidden runway cost in materials.

IIIScale

Can it be made, and at what margin?

  1. 6.Who manufactures at volume, and what does the cost curve look like?

    Lab-scale wins often die at kilogram and ton scale.

  2. 7.Do you sell material, license designs, or both? What are the margins on each?

    This decides whether the slow path to scale is your problem or your customer's.

IVDefensibility

What's still yours in five years?

  1. 8.If open materials models get as good as yours, what remains defensible?

    Assume the model is a commodity and see what's left.

  2. 9.Which incumbents could buy or build this, and why haven't they?

    Chemical majors move slowly until they don't.

If you're building at this edge, I want to hear from you.

Tell me what you're designing and who's waiting for it. I read every submission myself and connect the strongest founders with the investors and operators who should know them.

Next: Thesis 01 · Biomarker Expansion →
Raising now?

140 characters max