Reppo has landed a major strategic commitment — $20 million from Bolts Capital — to scale its prediction‑market protocol and build decentralized “Datanets” that turn staked human judgment into AI training and evaluation data, the Reppo Foundation announced on April 23.
What Reppo is building
- Reppo uses prediction markets to convert human beliefs into verifiable, incentive‑aligned data streams instead of relying solely on traditional labeling pipelines. Participants stake capital on predictions and are financially accountable for being wrong, which the project says yields sharper probability estimates and richer behavioral signals than conventional annotation or survey methods.
- The protocol supports multimodal inputs (text, images, audio, video) and organizes outputs into decentralized networks called Datanets — continuous, time‑stamped, scored datasets positioned for model training, evaluation, and fine‑tuning.
Why the funding matters
- Bolts Capital structured the $20 million as a strategic investment, signaling a long‑term bet on Reppo’s thesis and on the broader idea that prediction markets can solve AI’s training‑data bottleneck.
- The capital will accelerate protocol development, expand the Datanet ecosystem, and fund new prediction‑market primitives and tooling so AI teams can more easily plug Reppo‑derived data into their pipelines.
The stakes
- Reppo’s approach targets a core AI problem: noisy, biased, low‑signal datasets that can cap model performance even as compute increases. If markets can reliably produce higher‑quality training and evaluation data than traditional labeling services, Bolts’ commitment could be an early signal that crypto‑native prediction markets are evolving from speculative venues into foundational infrastructure for AI development.
Read more AI-generated news on: undefined/news
