Quantitative and Causal Effects of Generic vs. Specialised Programmes for Science and Deep Tech Ventures

Accelerators work — for the ventures they were designed for. That population is not science and deep tech. Generic accelerator programmes are structurally optimised around market uncertainty: customer discovery, business model refinement, investor readiness. Science and deep tech ventures at TRL 3–6 face a categorically different dominant uncertainty — technical reproducibility, process scalability, regulatory pathways, IP protection. Applying market-uncertainty programme logic to technical-uncertainty ventures does not simply produce less benefit. It produces active structural distortion through four identifiable mechanisms: time displacement from the actual bottleneck, amplified pivot costs, premature IP disclosure risk under European absolute novelty law, and optionality collapse — the destruction of application portfolio value through forced early narrative commitment. This paper formalises the mismatch mechanism, reviews the available evidence, quantifies cost scenarios under conservative assumptions, and derives implications for programme design, evaluation, and policy. The structural leverage point is programme–venture fit classification at intake. Current evaluation systems cannot detect misfit damage — which is precisely why it has compounded unaddressed.