What. I thought alcohol was good for you

This is a counterintuitive pattern in how shared UTXO systems (Spark, Ark, statechains, even Lightning to some degree) are being adopted. The usual assumption: Moving to lower-trust systems also means moving to more decentralized ones. Bitcoin base layer is the canonical example — minimal trust, thousands of nodes. Custodial services are the opposite — high trust, but at least there are many of them (countless exchanges, wallets, banks worldwide), so failure or capture of any one doesn't affect everyone. What's actually happening: The custodians and trusted intermediaries — which are individually trusted but collectively distributed across many independent entities — are migrating onto shared UTXO protocols. Those protocols do reduce trust (operators typically can't steal, only censor or refuse to cooperate on exit). But the operator set for any given shared UTXO system is tiny. Spark has a handful of SOs. Ark has few ASPs. Each system tends toward 1–N operators, not thousands. The paradox: You'd expect the trust/decentralization axes to move together. Instead, the industry is trading "many trusted entities" for "few less-trusted entities." Trust went down — good. Operator count also went down — not what the decentralization story usually promises. The implicit worry is the systemic risk: if most NZ/AU bitcoin activity ends up flowing through two or three shared UTXO operators globally, censorship, regulatory capture, or operational failure at any one of them affects a huge swath of users at once — even if none of them can outright steal funds. The trust model improved on one dimension while the concentration risk got worse on another.