A letter to founders

Every technology now taken for granted was once unproven. The proof came later. What came first was a person who had noticed something and could not yet say what it was worth: a material behaving oddly, a pathway that explained more than it should, an illness that ignored its own textbook. Sometimes it is just a calculation nobody had bothered to run.

That is when we want to meet you. Plenty of money shows up once the data is clean and the round is oversubscribed. We would rather be there while the idea is still mostly conviction, and your hardest problem is getting the next experiment to run.

KdT was founded to back that kind of work. A great deal of engineering talent has gone into moving information around and making it easier to act on. That work is hard, but it runs on a forgiving clock, where a mistake costs an afternoon. We were built for the layer underneath. A molecule has to behave inside a body, a catalyst has to survive its fiftieth cycle, a material has to hold at temperature. None of it can be patched so simply after launch.

We incubate and invest globally from pre-seed through Series A in companies whose advantage is scientific rather than commercial. That advantage might live in a reactor or a clinic, a fab or a model. When someone is first to work at that layer, what they get is rarely just a better product in an existing market. It is often a rearchitecting of an industry, a new treatment paradigm with clear benefit, or a model that turns a decade of experiments into a week of compute – and those all take years to stop looking unreasonable.

Working with us means a small number of companies and an unusual amount of attention. We read the data, all of it. Expect us to argue with your assumptions, because we would rather be wrong with you in a room than right about you afterward. The introductions you get are the ones we would make for ourselves. And when experiments fail – which they inevitably will – we will be by your side. If science were so easy, it would be called “search”, and not “re-search”. We understand that.

Capital is not the scarce input in frontier science. Judgment is, and so is the patience to stay while the data is messy and the story gets complicated.

If you are building something that only works if the science is real, we would like to hear from you.