YZI Labs has doubled down on Predict.fun, adding fresh capital to the on‑chain prediction market after the protocol graduated from the incubator’s EASY Residency Season 2. The follow‑on round brings in Susquehanna Crypto — the digital‑asset arm of quantitative trading giant Susquehanna International Group — marking a notable bit of institutional validation for prediction markets built on-chain.
Predict.fun, which YZI showcased during Binance Blockchain Week, combines DeFi yield, self‑custody and a gasless UX to turn event bets into yield‑generating positions. Instead of sitting idle, user collateral is routed into DeFi strategies while markets remain open, effectively making predictions a yield‑bearing primitive rather than simple wagers.
Since launching in December 2025, Predict.fun has processed more than 4 million orders and surpassed $1.8 billion in cumulative trading volume, according to figures shared by YZI Labs and third‑party reporting. Trust Wallet — which recently integrated the protocol — reports roughly $1.7 billion in volume across about 125,000 users and 3.7 million transactions on BNB Chain, underscoring rapid adoption on that network.
Built as a self‑custodial app on BNB Chain and transacting in USDT, Predict.fun resolves markets via a hybrid model: AI‑assisted proposals, human verification against reputable data sources, and UMA’s Optimistic Oracle. YZI Labs says this combination balances automation with accountability for high‑stakes outcomes.
Susquehanna Crypto’s participation links Predict.fun to one of the world’s largest options and ETF market makers, boosting the platform’s credibility as prediction protocols compete to become institutional‑grade trading venues. YZI also notes that Predict.fun is a standout in its EASY Residency network and a core beneficiary of the firm’s $1 billion Builder Fund for the BNB ecosystem.
For YZI Labs, the investment fits a broader strategy of backing projects at the intersection of Web3 and AI, where on‑chain markets can serve both retail traders and as structured data feeds for advanced models. If Predict.fun can sustain multi‑billion‑dollar volumes while keeping slippage low and resolutions reliable, the protocol is well positioned to challenge centralized prediction venues and emerging rivals across other L1s and L2s.
Read more AI-generated news on: undefined/news
