Problem

Prediction markets are constrained by the questions exchanges choose to list. Users begin with beliefs about future events, while exchanges require hand-written contracts with fixed wording, rules, and settlement criteria; as a result, many valuable, tradeable questions never exist in a usable market. ValCtrl points to Kalshi’s roughly 8,000 live questions and the fact that about 80% of its volume is sports as evidence of the gap: outside sports, informed trading can dominate, order books remain thin, and the exchange model becomes economically difficult to sustain.

The killer use case is a user who has a specific, public, resolvable belief—about geopolitics, politics, markets, deals, rates, or another future event—but cannot find an exact contract to trade. ValCtrl aims to turn that otherwise unpriced belief into a market and, in doing so, make prediction-market information useful beyond the heavily subsidized sports segment.

Product / Service

ValCtrl is building an intelligent prediction market in which a user types any public, resolvable belief about the future. The system structures the belief, finds related market signals, evaluates basis risk, prices the claim, and moves toward model-priced liquidity in which ValCtrl can take the other side subject to risk and compliance controls. Its initial live product, Kassandre, is a belief-search interface that maps a typed question to existing markets and exposes gaps between them.

The market itself is fully built on ValCtrl’s internal testnet, while public launch is gated on regulatory clearance. The intended business model combines per-trade fees with subscriptions to proprietary cross-market signals for insurers, funds, corporate-development teams, and supply chains; those signals are generated from trades and are intended to sharpen pricing and help fund winning payouts.

Market

ValCtrl competes in prediction markets and the emerging “Prediction Finance” category, positioning itself against listed-contract exchanges such as Kalshi by making the market itself responsive to user beliefs rather than limiting users to prewritten questions. The company was founded in 2025 by Sarth Garg and Gaurav Paliwal, is a Spring 2026 Y Combinator company, and is based in New York City with a two-person team. PitchBook reports $500,000 of funding and identifies Y Combinator as an investor.

Traction is currently product and infrastructure traction rather than demonstrated public-market revenue. Kassandre’s search product is live, with the company reporting 2.4-times daily belief-query growth from its first week, about 10,000 markets mapped, roughly 260 daily queries, and more than 100,000 correlations mapped on its internal testnet. Because the tradable market remains gated on regulatory clearance and the available evidence does not report customers or revenue, ValCtrl is best characterized as pre-public-launch rather than an established revenue-generating exchange.