The human cell cannot reliably tell a cure from an infection. When a gene therapy delivers new DNA into a patient, the cell often reads that therapeutic code the way it reads a virus, as something to be neutralised, and switches it off using a defence mechanism called the HUSH complex. The therapy works in the trial, then quietly loses power in the patient, which is one of the most expensive unsolved problems in modern medicine. Houdini Bio, a University of Cambridge spinout, emerged from stealth this month with about £1.5 million to make therapeutic DNA invisible to that defence.

The funding, roughly €1.7 million, combines non-dilutive grants with an oversubscribed pre-seed round led by SCVC, the venture arm of Science Creates, with participation from Deep Science Ventures and Cambridge Enterprise VC. All three are new investors, it is the company’s first public round, and no valuation was attached. The cheque is small by the standards of the field it is entering, which is part of what makes the claim attached to it worth examining.

A PhD discovery, turned into a platform

The science traces back to one laboratory finding. The HUSH complex was identified at the University of Cambridge by Paul Lehner, who worked out how the body silences foreign genetic material. Building on that during his own PhD, Jonathan Cohen-Gold discovered a set of molecular rules that let certain DNA sequences slip past the lock. Houdini Bio is the attempt to turn that insight into infrastructure: a machine-learning platform that reads a therapy’s genetic sequence before it ever reaches a patient and redesigns it to escape HUSH while keeping its medical function intact. The company calls the result a kind of genetic camouflage, and says its preclinical data show gene expression boosted by more than tenfold over current state-of-the-art methods, with technical milestones hit in a third of the time the team had projected.

Cohen-Gold runs the company as CEO alongside co-founder Lee Dunham, the chief business officer, who spent over a decade working with more than 100 cell and gene therapy companies, including as a director of business development at the Cell and Gene Therapy Catapult. That pairing, a scientist who found the mechanism and an operator who has watched promising therapies stall on exactly this problem, is the shape investors tend to want in a deep-tech spinout. The round was led by Harry Destecroix, the SCVC managing partner who founded the Bristol spinout Ziylo and sold it to Novo Nordisk for up to $800 million, which is the sort of track record that lends a tiny pre-seed more weight than its size suggests.

A small cheque against a large, crowded problem

The market underneath the science is genuinely big. The global gene therapy market is forecast to grow from around $10 billion in 2026 to roughly $26 billion by 2031, and durability, the question of whether a therapy keeps working once it is inside the body, is one of the field’s costliest open challenges. A CAR-T therapy that re-engineers a patient’s immune cells to fight cancer can lose strength within weeks as the inserted gene is silenced. If Houdini Bio’s approach keeps such a therapy active longer at lower doses, the value is obvious, both to patients and to the manufacturers paying enormous sums to produce these treatments.

The reasons for caution are equally clear, and they start with the size and stage of everything involved. This is a £1.5 million pre-seed behind a preclinical platform, and the headline figure that sells it, the tenfold improvement in gene expression, rests on the company’s own data rather than independent validation. Durability is also a popular target right now, which means competition. nChroma Bio, formed when Chroma Medicine merged with Nvelop Therapeutics in late 2024, raised $75 million from a syndicate including Cormorant, ARCH Venture Partners and Atlas Venture; Dyno Therapeutics, in Cambridge, Massachusetts, is using AI on the delivery side of the same problem. Houdini Bio’s distinct bet is that the answer lives in the DNA sequence itself rather than in the viral shell that carries it or the disease gene it targets. That focus is its edge and its exposure: if the sequence-level approach is right, it is a platform many therapies could license; if a rival cracks durability another way first, a small spinout has the least room to absorb being second.

The company graduated from the third cohort of Science Creates’ Engineering Biology Accelerator, and says the round buys it the runway to expand the team and strike co-development partnerships with pharmaceutical companies looking to fold anti-silencing technology into their pipelines. That, not a drug of its own, is the near-term business: selling the disguise rather than the medicine.

It is a fitting plan for a company named after an escape artist. The trick a human cell plays on a cure, mistaking the medicine for an invader and quietly shutting it down, has cost the gene-therapy industry years and fortunes. Houdini Bio is wagering that the way out is not a stronger dose or a cleverer delivery vehicle but a better disguise, written into the genetic code itself. Whether the disguise holds once it leaves the lab and meets a real patient is the whole company, and £1.5 million is the budget for finding out.

Filed by eu-tailwind. Independent reporting on the rise of European tech.