As the cost of computational science (modeling, simulation, discovery, publishing) trends toward zero thanks to AI, and as research outputs become increasingly modular, agentic, and composable, we can expect a Cambrian explosion of scientific IP to emerge organically. Yet real-world wet lab work remains capital-intensive. Agentic AI will speed up discovery by automating hypothesis testing, optimizing experiments + iterating protocols at unprecedented scale. This empowers a new science “idea guy” to bootstrap breakthroughs without massive upfront teams. That said, how will these outputs find funding? Every medium for new scientific IP has trended toward lower barriers: arXiv preprints, open-source repos, citizen science notebooks, short-form research tweets/threads… leading naturally to onchain tokenized assets (decentralized IP, agent-minted research DAOs, tokenized datasets/protocols) as the endgame Onchain science assets are low-cost to launch, global from day one, with programmable scarcity + provenance sans gatekeepers. Mints serve as a capital formation mechanism to crowdfund validation, wet-lab scaling, or replication. This will become the fastest, lowest-overhead way to bootstrap internet-native science @Molecule_sci tokenisation frameworks + @BioProtocol scientific AI will turbocharge this flywheel