Cusp
The capital markets layer for prediction markets.
Cusp is risk, credit, settlement, and liquidation infrastructure for event-driven assets: the positions traded on prediction markets and, more generally, any short-maturity claim that resolves against a contractual source and settles after the fact.
The problem
Prediction markets have reached scale, but their positions remain financially inert. A claim can be bought, held, sold, or redeemed at resolution, and nothing else. There is no credit against it, no yield on it, no financing for the gap between resolution and payout, no shared standard a third-party lender could underwrite against, and no liquidation mechanism suited to how these assets actually behave.
Every comparable on-chain asset class acquired this infrastructure within months of reaching scale. Event contracts have not, and the reason is structural. An event position is bounded between 0 and 1, reprices in discontinuous jumps as information arrives, loses liquidity at exactly the moments a liquidator would need it, attracts informed borrowers by construction, and terminates in a binary payoff that can erase collateral value in a single block. Each of those properties breaks an assumption that conventional overcollateralized lending depends on. Bolting a generic lending market onto this collateral does not produce a credit market; it produces an insolvency mechanism with extra steps.
The vision
Assets become asset classes when capital markets form around them: when they can be measured, described in a common language, financed, structured, and unwound. Cusp's thesis is that event-driven assets do not need a lending product adapted to them; they need a capital market constructed for them.
One principle organizes the whole design. For collateral that reprices in jumps and loses liquidity when it matters, solvency cannot be recovered after a drawdown. It has to be arranged beforehand, through conservative valuation, state-contingent pricing, hard limits, and explicit capital structure.
The market
Aggregate prediction-market volume grew from roughly $3B in 2023 to about $16B in 2024 and more than $60B in 2025, with early-2026 monthly volume implying an annual run rate several times larger. Published sell-side projections place the sector near $1 trillion in annual volume by 2030. Institutional capital has followed: the two largest venues carry private valuations near $9B and $11B, the exchange operator that owns the New York Stock Exchange has committed up to $2B to the sector, and survey evidence suggests nearly half of proprietary trading firms are evaluating event-contract strategies. Figures are approximate and vary by methodology; the order of magnitude and the growth rate do not. None of the surrounding capital market exists yet. That is the gap Cusp fills.
Core components
Cusp ships as seven connected modules.
Risk engine
Reads live order books, the venue's authoritative trade record, and venue metadata, and publishes per market an eligibility verdict, a conservative collateral value, limits, pricing, exposure caps, and a liquidation mode.
Position standard
A published, versioned metadata object that makes every supported position legible to vaults, liquidators, external lenders, and builders. Every score is falsifiable against the outcome that realized.
Instant Redeem
Settlement financing. Cusp purchases resolved winning claims during the settlement gap, so holders get liquidity immediately while the protocol collects face value from the venue.
Tranched vaults
Senior capital is protected by structure; staked first-loss capital absorbs losses first and earns the levered residual. Idle balances earn a base money-market rate from the first block.
Controlled credit
Borrowing against eligible live positions, written in short repriced terms, with capacity that decays toward resolution and a gate that stops new credit entirely near it.
Liquidation auctions
Descending-price auctions among bonded specialists, opening at the conservative mark, settling atomically, with surplus above debt and costs returned to the borrower.
Calibration and transparency
Score history, diagnostics, parameter changes with reasons, and the complete liquidation log, all public.
What makes the design different
- Measured from the right data. Trade-signed inputs come from the venue's authoritative record, because public display feeds get direction wrong on a large share of activity.
- Conservative by construction. Collateral is valued at the most pessimistic of several prices, never the last trade, with stress haircuts on top. No defensible stress value means no credit.
- Priced per state, per epoch. Credit terms are re-derived from current market state at every rollover instead of from a single utilization curve.
- Settlement financing as a first-class product. Resolved claims are purchased as receivables, an exposure with no price risk and no adverse selection, priced with conservatively inflated failure assumptions.
- A provable launch posture. At launch, total credit is bounded by first-loss capital, so senior depositors are protected by arithmetic rather than by estimation, and the protocol scales only as fast as its published calibration record allows.
- Auditable by design. Everything the system decides is published and checkable against realized outcomes.
Resources
- Whitepaper: Cusp v1, the formal treatment of all of the above (models, bounds, and mechanism analysis).
- Documentation: high-level guides to each component.
- Contact: research@cusp.fi
Disclaimer
This document describes protocol design. Nothing here is investment, legal, or tax advice. Market figures are compiled from venue disclosures and public reporting; the 2030 figure is a published sell-side projection, not a Cusp forecast.