Should you build or use a B2B platform?

The first decision for any operator adding prediction markets is infrastructure approach:

ApproachTimelineCostBest for
Build from scratch12–18 monthsHigh (engineering + trading + compliance)Tier-1 operators with existing prediction exchange teams
B2B platform8 weeksPlatform fee + rev-shareMost operators — casino-first, startup, or sportsbook wanting new vertical
Widget-only API2–4 weeksLowTesting only — limited liquidity, no cross-sell, no operator reporting

For most licensed operators, a B2B platform is the correct choice: no trading desk required, pre-integrated compliance for key jurisdictions, and casino cross-sell wired from day one. Building from scratch is only cost-effective for operators expecting $50M+ annual GGR from prediction markets within 24 months.

Step 1: Choose your liquidity sources

Prediction market platforms aggregate liquidity from external venues — you do not build your own order book. The main venues as of mid-2026:

  • Polymarket (live): Largest decentralised prediction exchange. $20B+ monthly volume in Jan 2026 (TRM Labs). USDC-settled on Polygon. Broad sports, politics, finance, entertainment coverage.
  • Kalshi (live, Q3 2026 B2B): US-regulated (CFTC). $43B notional volume in 2025, 89% from sports (KPMG 2026). The preferred regulated B2B liquidity source.
  • Manifold (Q4 2026): Play-money market with large coverage that transitions to real-money settlement. Good for long-tail event coverage.
  • On-chain venues (2027+): Decentralised markets with automated settlement via oracles.

Key principle: use a multi-source aggregator. This normalises all venues into one operator-facing order book — one outcome per event, one settlement path. When new venues launch, they're added inside the aggregator without touching your integration.

Step 2: Compliance by jurisdiction

Jurisdiction choice determines how quickly you can launch and what compliance infrastructure you need:

JurisdictionStatusKey advantageComplexity
Curaçao✅ Live nowRuns under existing Curaçao gaming licence. No new entity. Fastest path.Low
MGA (Malta)🟡 In reviewEU regulated. Same licence as casino. High credibility market.Medium
Anjouan✅ Live nowCrypto-native track. USDC/USDT settlement. L1+L2 rails.Low
Tobique / Kahnawake✅ SupportedNorth American operators. Indigenous gaming licence compatible.Low
US (CFTC)🔴 ComplexRegulated access to US players via DCM/SEF path.Very high
UK / EU Tier-1⏳ No framework yetFuture: large markets, high CPAs.Pending regulation

Recommended for first deployment: Curaçao. It's the fastest path, runs under your existing licence, and the Turbo Stars platform has live multi-operator deployments there.

Step 3: Wallet and cross-sell setup

This is where the business case is built. Prediction markets are thin-margin on their own. The value comes from cross-sell economics:

  • Same wallet: prediction markets, casino and sportsbook share one player account, one KYC flow, one balance. A prediction-market bet leads straight to casino — no re-deposit, no re-KYC.
  • Week-1 cross-sell: 41% of prediction-market players placed a casino bet in week one (Turbo Stars Curaçao launch, Q2 2026).
  • LTV uplift: cross-sell cohort shows +28% LTV vs casino-only baseline.
  • Sport alignment: 70% of prediction-market bets are sport-aligned. Feed the sportsbook retention loop with the same audience.

Wire the cross-sell at the wallet level, not the product level. When a player's prediction market session ends, the casino should be one tap away — same session, same balance, no friction.

Step 4: Risk configuration

Prediction markets don't require a trading desk but they do require risk configuration:

  • Exposure limits per market: set maximum operator exposure before circuit-breaker fires (auto-suspend or re-route to different venue)
  • Market suspension triggers: late-breaking news, data feed outages, disputed outcomes
  • Settlement workflow: automated for standard outcomes; escalation path for disputed events
  • Jurisdiction reporting: regulatory dashboards, AML transaction monitoring, RG flag thresholds

Step 5: 8-week launch timeline

WeekMilestone
1–2Technical integration: platform API, wallet connection, KYC flow
3–4Compliance setup: RG tools, reporting, jurisdiction configuration
5–6Casino cross-sell wiring: triggers, bonus engine, player journey
6–7QA, risk configuration, internal testing, soft launch (limited markets)
8Full launch: sport-aligned markets live, cross-sell active

What to expect in the first 90 days

  • Weeks 1–4: prediction-market audience acquisition. Focus on sport-aligned events (football, basketball, US sports).
  • Week 1: 41%+ of depositing players cross-sell to casino (partner data baseline).
  • Month 2–3: sportsbook retention loop activates — prediction-market players re-engage with sportsbook on the same events.
  • Month 3–6: measure combined LTV vs casino-only cohort. Target: +25–35% uplift.
  • Month 5–6: full media-buy payback on prediction-market acquisition (170% 12-month ROI, partner data).

Common mistakes to avoid

  • Launching without casino cross-sell: prediction markets alone are thin-margin. The business case requires the casino cross-sell wired from day one.
  • Starting with financial/political markets: 70% of volume is sport-aligned. Launch with sports first — it matches your existing audience.
  • Building from scratch for a test: 12–18 months and a trading desk for what a B2B platform delivers in 8 weeks without one.
  • Single-venue dependency: connect to at least two liquidity sources from launch (Polymarket + Kalshi). Venue outages are real.
  • Ignoring jurisdiction timeline: US (CFTC path) takes 12–24 months. Start with Curaçao, build the operational playbook, expand to regulated markets later.

Continue reading: Prediction markets operator solution — the full Turbo Stars launch package. Prediction markets vs sportsbook — the business model comparison.

Frequently asked questions

How long does it take to launch prediction markets?

8 weeks with a B2B prediction markets platform. The Turbo Stars platform completed its first Curaçao deployment in 8 weeks, with casino cross-sell active by week six. Building from scratch takes 12–18 months.

Do I need a new licence to offer prediction markets?

For Curaçao, Malta (MGA), Anjouan, Tobique, and Kahnawake: prediction markets can run under your existing licence. For the US: you need CFTC registration as a Designated Contract Market or Swap Execution Facility — a separate, complex process. UK and most Tier-1 EU jurisdictions do not yet have an explicit prediction markets framework.

Do I need a trading desk to run prediction markets?

No. Unlike a sportsbook where you price odds and hold risk, prediction markets run on external order books (Polymarket, Kalshi, Manifold). The platform aggregates these as liquidity sources. You earn platform fees without pricing markets. No trading desk, no market-making staff required.

What ROI should I expect from prediction markets?

Prediction markets are thin-margin (1–5% fee on volume) but close via casino cross-sell. In Turbo Stars partner data (Curaçao, Q2 2026): 41% of prediction-market players placed a casino bet in week one. Casino margin of 25–40% on that cohort generates +28% LTV uplift vs casino-only. Combined 12-month ROI on media-buy campaigns: 170%.

Which prediction market venues can I connect to?

Polymarket is live (the largest decentralised prediction exchange, $20B+ monthly volume). Kalshi is live in Q3 2026 (US-regulated, 89% sports revenue). Manifold is planned for Q4 2026. On-chain venues are on the roadmap for 2027. A multi-source aggregator adds venues without requiring operator re-integration.