Mental models for deeptech allocating

When Sam Cash went out to raise a first fund three and a half years ago, the pitch was an argument. Look at these facts, it ran, and you have to conclude that these industries and technologies will be very interesting. Investors recoiled. That was far too much work, they said. They had been told AI was interesting and had no time to take the rest to their investment committee. The lesson, Sam said, is that "the world of capital is often very lazy." It wants to underwrite a theme it already sees in the market.

Sam, who runs Entropy Industrial Capital, backs very technical founders at pre-seed. Kyle McEneaney of The Schmidt Family Foundation uses philanthropic capital to get new energy technologies to their first commercial deployment. Their question was what allocators need to judge deep tech. The answers kept coming back to translation.

Deep tech and deep tech-ish

Sam defines deep tech as novel or low-level technology, deep in the stack and often hardware, carrying both high technology risk and high market risk. Complex businesses that use existing technology in known markets are what Sam calls deep tech-ish. A pre-seed investor preferred a test that moves with the sector. A technology never before used in construction is deep tech in construction. Another investor agreed that novelty is the only workable definition but found market risk lower than technology risk in their own portfolio. Sam said a new technology often creates a new market, and that carries its own risk. Three or four years ago nobody predicted that AI inference compute would be bought by the gigawatt instead of by the rack.

A narrow Venn diagram

Kyle said deep technical skill looks like a prerequisite for a fund manager but rarely comes in the same person as the skill to raise money, so choosing a team is the hardest part. Sam called the overlap between the highly technical and the good communicators very narrow, for founders and GPs alike, and admitted wanting to talk engineering to allocators who care more about impact and are not equipped to judge the science risk a fund is taking. A founder described the same gap from the start-up side. A company has to sell a compelling vision while keeping a realistic roadmap, and investors who are not technical follow the narrative until fundraising gets hard. Then it bites. In the US, Sam said, Founders Fund concluded some years ago that the technical operator and the outward-facing evangelist can be two different people.

An investor with a physics background said they can follow a colleague's biology deals at a high level but cannot see the nuances between companies. Investors who cannot tell companies apart put money where the hype is. Kyle said the foundation is an unusual LP. Its principal is a deep-tech founder and investor with a network of experts to validate deals, which most LPs lack. Not everybody can be an expert in every topic, Kyle said. Sam wanted more domain experts to become GPs or venture partners.

The appetite for green premium in most of the world is zero at this point.

Labels, and who leads

About 80% of European funds now call themselves deep tech, one investor said, and in five years that will bring the same disappointment that followed climate tech. Sam said generalists back legible founders and familiar technology in whatever theme their investment committee can grasp. That was climate four or five years ago and is robotics, manufacturing and defence now. A founder making nanomaterials to replace critical minerals had lived through the relabelling. The company was climate four years ago and defence two years ago, and the new label scared some investors off. Sam's advice was to meet capital halfway. A climate company might do better calling itself a resilience company.

Whether good companies still find money divided the table. Kyle said the US climate-venture boom of three or four years ago rested on policy the new administration has rescinded, and even then many companies hit a wall venture could not fix. They needed debt for a first plant, such as a novel electrolyser or a low-carbon cement line. Lenders do not take technology risk and ask for 10 or 20 plants first, while venture investors will not put half a billion dollars into finding out whether one works. Sometimes a small technology-risk insurance policy is enough to bring private credit in.

A commercialisation-stage investor said the best companies do find capital, and what the ecosystem lacks is corporate partners, customers and joint ventures. "The appetite for green premium in most of the world is zero at this point," they said. Kyle disagreed. Strong climate companies are raising oversubscribed rounds that nobody will lead or price, and some may fail with the money sitting there. A corporate venture investor in agri-food saw the same at pre-seed, where longer exits have pushed investors later. Filling that gap takes longer horizons or lower multiples, or both, and capital from governments, corporates and family offices with different expectations.

Until more scientists sit on the other side of the table, Sam does the translating. Entropy spends 12 to 24 months with its pre-seed founders, packaging them into something later-stage capital can read without taking on the science risk.

This Ripple was hosted by Sam Cash (Entropy Industrial Capital) and Kyle McEneaney (The Schmidt Family Foundation) at The Drop 2026 on 16 September.