What if the necessities we take most for granted are the ones changing fastest? Energy, food and care reach us through systems designed for another technological era.
Now their physical foundations are being rebuilt. In the power plant, the paddock and the clinic, founders are applying advances in intelligence, biology and engineering within markets governed by rules that predate those capabilities. Technology enters these markets by earning permission, because they are systems society cannot afford to fail.
We lead seed rounds for founders rebuilding how populations are powered, fed and healed.
Necessity is the thesis
A model reads a scan, steers a machine across a paddock and holds a grid inside its limits. Every output is sold under a rule written before the capability existed.
A model trained on physical behaviour acts on a thing with mass: a beam of X-rays, a boom of spray nozzles, a bank of inverters.
In an open market a model compresses the technical lead. It cannot compress the evidence behind an approved product, because clinical validation, regulatory history and real-world performance accumulate only over time.
Existing evidence supports the requirement it was collected for, so a company designed for the new requirement is already gathering the data that one demands. hmm looks for that transition: the point at which regulation changes what must be proven and a younger company is already proving it.
Reaching a regulatory milestone costs less than it did, from discovery through to submission, so more companies now reach the bar than reached it before.
Permission does two different jobs.
A necessity system runs to as many as ten stages. The gate marks the stage whose value is hardest to compete away, which is why hmm's capital points there.
An approval is issued against a named site and the process running on it. Moving assembly to a building outside the certificate restarts three filings: the new site registers with the regulator as a manufacturing location, the conformity investigation, in which the regulator examines the process on that site against the standard, runs again there, and the change clears before the approval reads on the new line. Those three filings are the qualification clock, and an acquirer buys the approval because buying it keeps the clock where it stands. An entry changes hands administratively, and the clinical validation, the regulatory history and the real-world performance behind it stay where they were built, because a regulator grants an entry and time accumulates a record.
The data that locates a company also shows what its approval will cost, in time and in evidence, which is why every deal is read four ways against the instrument that gates it: which instrument it must clear, what evidence that demands, how far the field has already moved, and where the company sits on that path today. An AI diagnostic clears Japan's conditional early-approval route, or Australia's software-as-a-medical-device rule, on clinical evidence. Annual United States authorisations of AI-enabled devices rose from a maximum of six a year before 2016 to 331 in 2025, on the counts Singh and others published in 2025 and Golshani and Joseph in 2026.
Companies whose product cannot be sold without an approval reach a liquidity event at 17.73%, against 6.08% where no approval gates the sale, measured on 12,664 companies across the three markets at an odds ratio of 3.33. The tier between them, where a buyer sits under a compliance obligation, reaches 5.82% and sits with the unregulated rate.
Output per venture dollar, against the capital already bidding.
Both halves carry weight, because a screening field earns nothing where twenty funds bid on the same company.
Three markets, one distributed experiment.
Entry is priced by the local seed market and the exit is priced by the acquirer's market, which for these companies is international, so the two prices are set by different buyers.
How do we reach these founders before a round exists?
Source before the first round
These companies enter conventional venture networks late, if at all, and many founders are still on the technical problem when hmm first approaches them. So the fund maps a market, identifies the companies inside it and contacts the teams before a formal raise begins.
Build the commercial side
Technical capability is often already present, and the unresolved questions are commercial: who buys, what budget pays, how large the reachable market is and how the company reaches its first customers. hmm works through those questions with the founders.
Enter locally, scale internationally
Exit route differs by market, and limited local capital continues to shape entry pricing, so hmm invests before international capital becomes the price setter. It then helps founders build toward the markets where their customers and eventual acquirers operate.
The first conversation is a research conversation about your market: what is being built there, which rule gates it and who pays once it clears.
It is the one thing we ask for, and it is where the map starts. wt@hmm.ventures