Landscape
Emerging platforms, mapped on the two axes that matter: how wide they go, and where they settle.
Click a quadrant to isolate it · click a platform for details · from the Predicted Q1 2026 report
Click a quadrant to isolate it, or a platform for details.
General: wide scope, many categories
Specific: limited to one theme or vertical
Onchain: trades and outcomes managed by smart contracts or decentralized ledgers
Offchain: traditional, centralized systems for trading and settlement
Four business layers: distribution on top, brokerage under it, then the exchange, then clearing. The licences at the bottom are the moat, and in 2026 everyone with distribution is buying their way down.
Distribution & interface apps, wallets, DEX front-ends | Rents Robinhood, Webull, Phantom, Jupiter, DFlow | Owns own app + wallet integrations | Owns the Robinhood app | Owns IBKR platform | Owns Crypto.com app | Owns FanDuel app + Predicts | Owns DraftKings app | Owns app — surfaces Kalshi today | Owns Underdog app | — institutional / broker network |
Brokerage (FCM) customer accounts, order routing | — direct app + partner FCMs | — n/a — direct access | Owns Robinhood Derivatives | Owns IBKR FCM | — integrated in-app | Building own FCM filed Apr 2026 | Building FCM application pending | Owns Coinbase Financial Markets | — | — |
Exchange (DCM) the order book and matching | Owns DCM | Owns QCX (DCM) + onchain CLOB globally | Owns Rothera live (ex-LedgerX/MIAXdx) | Owns ForecastEx (DCM) | Owns DCM | Rents CME JV — CME owns 51% | Owns Railbird DCM (Oct 2025) | Rents runs on Kalshi's product for PMs | — via partners today | Owns Cboe futures exchange |
Clearing (DCO) settlement and guarantees | Owns DCO | Owns QC Clearing (DCO) | Owns Rothera DCO — ~$11.6M/day (Artemis) | Owns ForecastEx (DCO) | Owns DCO | Rents CME Clearing | Rents CDNA today | Building via Kalshi today · buying The Clearing Co (CFTC pending) | Owns Aristotle's DCO, acquired | Rents OCC |
Wallets and apps build on top; exchanges and clearinghouses sit at the bottom; the licences (DCM to run markets, DCO to clear them) are the moat. Click a company to read its stack strategy. The pattern of 2026: everyone with distribution is buying the bottom of the stack.
Distribution-only renters not shown as columns: Fanatics, OG, PredictIt, ProphetX, Sporttrade — and Novig, which deliberately skipped a DCO. See the clearing map below for where they settle.
Every brand needs somewhere to clear, and there are far fewer clearinghouses than storefronts. This is the consolidation at the bottom of the stack.
Four brands clear through Crypto.com's CDNA (ex-Nadex): Crypto.com/OG, Fanatics, DraftKings, and FanDuel's new leg. The former Nadex quietly became the sector's clearing utility.
Everyone else is racing to own a clearinghouse. Novig's Ludlow Exchange has no DCO at all — it self-settles via 100% collateral, skipping the clearing layer the whole sector is fighting over.
Hover to trace who clears where. Gold = a Q2 2026 event; the story of the quarter is rent → own (DraftKings, Robinhood and Kraken all moving onto their own clearing). Verified against CFTC records · sourcing in Q2_MA_RESEARCH.md.
Prediction markets are multichain, and lopsided: Solana leads by platform count, off-chain venues hold the deepest liquidity, and whole mechanism-and-chain combinations sit empty.
Platforms deploying on each chain; a multichain platform counts once per chain. From PM Atlas's public dataset of 125 platforms.
Darker means more venues; a dashed cell is a market no one has built yet. The lesson borrowed from PM Atlas: the gaps are the opportunity.
| CLOB | AMM | LMSR | Parimutuel | |
|---|---|---|---|---|
| Solana | 6 | 7 | 1 | 8 |
| Base | 6 | 6 | 1 | 2 |
| Polygon | 9 | 2 | open | 2 |
| Arbitrum | open | 6 | 1 | 5 |
| BNB Chain | 4 | 4 | open | 1 |
| Hyperliquid | 5 | 1 | open | open |
| Off-chain | 20 | 10 | 1 | 1 |
Two things jump out: LMSR— Augur's original mechanism — is nearly extinct, at most one venue per chain; and Arbitrum has no order-book venue despite hosting AMM and parimutuel ones. Order books cluster off-chain and on Polygon; the newer chains lean AMM and parimutuel.