DraftKings Predictions: Event Contracts, Railbird Acquisition & Agent Trading Guide

DraftKings Predictions is the most important product to emerge from the convergence of sportsbooks and prediction markets. It is the first time a publicly traded, mass-market sportsbook operator has launched a full prediction market product – binary event contracts traded on an order book, regulated at the federal level, accessible to millions of existing DraftKings users. If you are building AI sports betting agents or automated trading systems that operate across wagering platforms, DraftKings Predictions is the product that forces you to rethink your architecture.

The significance is not just that DraftKings added a new feature. It is that DraftKings Predictions represents a structural bridge between two ecosystems that have operated independently for years: state-regulated sportsbooks where you bet against the house, and federally regulated event contract exchanges where traders set prices against each other. DraftKings now operates in both worlds simultaneously, and the implications for arbitrage, market-making, data analysis, and agent design are substantial.

This guide covers everything developers, quantitative traders, and agent builders need to know about DraftKings Predictions: how it works mechanically, the Railbird acquisition that made it possible, head-to-head comparisons against Polymarket and Kalshi, concrete arbitrage strategies between DraftKings Predictions and DraftKings Sportsbook, the current state of API access, and how to architect agents that are ready for the converged future.

If you are new to the distinction between sportsbooks and prediction markets, start with our sports betting vs. prediction markets guide. If you want the broader convergence picture, read the convergence analysis. This article goes deep on one product and why it matters.

What Are DraftKings Predictions?

DraftKings Predictions is a binary event contract platform. That description sounds simple, but every word in it matters, because it distinguishes the product from what most DraftKings users are familiar with.

Binary means every contract has exactly two outcomes: Yes or No. There are no point spreads, no over/unders, no multi-way moneylines. You either believe an event will happen, or you believe it will not.

Event contract means you are buying and selling a financial instrument, not placing a traditional bet. Each contract is priced between $0.00 and $1.00. If the event occurs, the contract settles at $1.00 and every holder receives one dollar per contract. If the event does not occur, the contract settles at $0.00 and holders receive nothing. The price at any given moment reflects the market’s collective estimate of the probability that the event will occur.

Platform means DraftKings Predictions is not a feature within the sportsbook – it is a separate product with its own regulatory framework, its own market mechanics, and its own settlement rules. It happens to live inside the DraftKings app and shares a wallet with the sportsbook, but structurally it is a different kind of business.

Here is the mechanics of a single trade. Say a contract asks: “Will the Federal Reserve cut interest rates at its June 2026 meeting?” The current market price is $0.62 Yes / $0.38 No. If you believe the probability is higher than 62%, you buy Yes contracts at $0.62 each. If the Fed does cut rates, your contracts settle at $1.00 and you profit $0.38 per contract (a 61.3% return). If the Fed does not cut rates, your contracts settle at $0.00 and you lose your $0.62 per contract. You can also sell your contracts before settlement if the price moves in your favor – or against you – and you want to exit the position.

This is fundamentally different from placing a bet at DraftKings Sportsbook. On the sportsbook, DraftKings sets the odds, takes the other side of your wager, and manages its own risk. On Predictions, DraftKings operates the exchange – it matches buyers and sellers, facilitates settlement, and takes a fee. The house does not have a directional position on whether the Fed cuts rates or not. The market participants set the price by trading with each other.

The event categories available on DraftKings Predictions extend well beyond sports. Current and past markets have covered:

• Politics: Presidential election outcomes, congressional races, gubernatorial elections, Supreme Court decisions
• Economics: Federal Reserve rate decisions, inflation data releases, jobs reports, GDP figures
• Entertainment: Award show winners (Oscars, Grammys, Emmys), TV show outcomes, box office performance
• Weather: Temperature records, hurricane landfalls, seasonal snowfall totals
• Sports-adjacent: MVP awards, draft picks, trade deadlines – events that are related to sports but not direct game outcomes
• Current events: Geopolitical developments, technology milestones, cultural moments

This breadth of coverage is what makes DraftKings Predictions strategically significant. Traditional sportsbooks are constrained to events that state gaming commissions approve for wagering, which is overwhelmingly sports. Prediction markets can list contracts on essentially anything measurable, and DraftKings Predictions inherits that flexibility through its federal regulatory framework.

The Railbird Acquisition

DraftKings Predictions did not emerge from an internal R&D project. It was built on an acquisition, and understanding what DraftKings acquired explains how the product works, why it is regulated the way it is, and where it is likely headed.

What Was Railbird?

Railbird was a startup founded with the specific goal of building a prediction market exchange for event contracts. Unlike most crypto-native prediction market startups that emerged around the same time, Railbird pursued the regulated path from the beginning. The company applied for and received designation as a CFTC-registered designated contract market (DCM) – the same regulatory classification held by Kalshi and major commodity futures exchanges like the CME.

The DCM designation is the critical asset. Obtaining it requires demonstrating to the CFTC that you can operate a fair, transparent, and compliant marketplace for derivatives trading. The application process involves detailed submissions on market surveillance, compliance procedures, financial safeguards, and technology infrastructure. It typically takes one to two years from application to approval, and the CFTC has granted only a handful of new DCM designations in the event contracts space.

Railbird’s technology stack was purpose-built for binary event contracts: an order matching engine, a contract specification framework for defining events and settlement criteria, user-facing trading interfaces, and compliance infrastructure for KYC/AML obligations under federal regulation. The team included engineers and compliance professionals with backgrounds in both fintech and regulated exchanges.

The Acquisition Timeline

DraftKings announced its acquisition of Railbird in late 2024. The strategic rationale was transparent: the 2024 presidential election had demonstrated massive consumer demand for event-based wagering, and platforms like Polymarket and Kalshi were capturing that demand while sportsbooks watched from the sidelines. DraftKings recognized that it could either build a prediction market product from scratch – a multi-year process that would require its own CFTC application – or acquire a company that already had the license, the technology, and the team.

The acquisition gave DraftKings three things:

1. A CFTC DCM license. This is the crown jewel. With Railbird’s DCM designation, DraftKings can operate a federally regulated event contract exchange without going through the years-long application process itself. The license allows DraftKings Predictions to list contracts on events that would not be permissible under state gaming commission rules.

2. A purpose-built technology stack. Railbird’s matching engine, contract framework, and compliance infrastructure provided the foundation for DraftKings Predictions. Rather than retrofitting its sportsbook technology to handle order-book trading, DraftKings could deploy a system that was designed for exchange-style markets from the ground up.

3. A team with regulatory expertise. Operating a CFTC-regulated exchange requires ongoing compliance with federal regulations that are materially different from state gaming commission oversight. Railbird’s team brought expertise in federal derivatives regulation, market surveillance, and the specific compliance obligations of a designated contract market.

Integration Into DraftKings

The integration strategy was to embed Predictions within the existing DraftKings app rather than operating it as a standalone platform. This was a deliberate decision that traded independence for distribution. A standalone Railbird exchange would have needed to acquire users from scratch. Embedded within DraftKings, Predictions has immediate access to millions of existing accounts, a shared wallet system that lets users move funds between sportsbook and prediction market seamlessly, and a familiar brand that reduces friction for first-time event contract traders.

The integration also created something no other platform has: a single app where a user can place a traditional sportsbook bet on an NFL game and trade an event contract on the Federal Reserve in the same session, from the same wallet. This coexistence is DraftKings Predictions’ most significant competitive advantage and its most interesting feature from an agent architecture perspective.

How DraftKings Predictions Works

Understanding the mechanics at a granular level matters for anyone building automated systems that might interact with this platform.

Contract Structure

Every DraftKings Predictions contract is a binary instrument with a defined event, resolution criteria, and expiration date.

• Event definition: A precisely worded question with an unambiguous Yes or No answer. Example: “Will Bitcoin close above $100,000 on June 30, 2026?”
• Resolution source: Each contract specifies the authoritative data source used to determine the outcome. This could be an official government agency (Bureau of Labor Statistics for jobs data), a recognized standards body (Associated Press for election calls), or a verifiable third-party source.
• Expiration: Every contract has a defined expiration date and time at which the outcome is determined and the contract settles.
• Settlement: Contracts that resolve Yes pay out $1.00 per contract. Contracts that resolve No pay out $0.00. Settlement typically occurs within hours of the resolution event, though complex events may take longer if the resolution source requires time to publish results.

The Order Book

DraftKings Predictions operates as an exchange, which means prices are set by an order book – the aggregation of all outstanding buy and sell orders from market participants.

Limit orders let you specify the price you are willing to pay. If you want to buy a Yes contract but only at $0.55, you submit a limit order at $0.55. It sits on the order book until another participant is willing to sell at that price.

Market orders execute immediately at the best available price. If the current best offer for Yes is $0.62, a market buy order fills at $0.62.

The bid-ask spread is the difference between the highest price a buyer is willing to pay and the lowest price a seller is willing to accept. Tight spreads indicate deep liquidity. Wide spreads indicate thin markets where you will face significant slippage on larger orders.

For developers accustomed to Polymarket’s CLOB or Kalshi’s order book, the DraftKings Predictions order book operates on the same fundamental principles. The difference is access: Polymarket and Kalshi provide APIs that let agents submit and manage orders programmatically. DraftKings Predictions, as of early 2026, does not.

Pricing and Implied Probability

Because Yes and No contracts for the same event must sum to $1.00 (one of them will pay out, the other will not), the market price of a Yes contract directly represents the market’s implied probability for that outcome. A Yes price of $0.72 means the market collectively estimates a 72% probability that the event will occur.

This is cleaner than sportsbook odds, which embed the vig. When DraftKings Sportsbook prices an event at -250 (implied probability 71.4%), the true market probability is lower – the vig inflates the implied probability. When DraftKings Predictions prices the same event at $0.68 Yes, that $0.68 is the market’s unvarnished estimate, though the platform fee on trading or settlement represents the exchange’s cost.

Fee Structure

DraftKings Predictions charges fees on transactions. The specific fee structure has evolved since launch, but the general model involves a per-contract fee on trades and/or a fee on settlement payouts. Unlike a sportsbook’s vig, which is embedded invisibly in the odds, exchange fees are explicit and disclosed. This transparency matters for agents that need to calculate expected value precisely – you can model the fee as a known cost rather than estimating it from overround analysis.

Event Categories in Detail

DraftKings Predictions has expanded its event catalog steadily since launch. The categories that see the most volume and liquidity are worth understanding for any developer evaluating the platform.

Political markets have been the highest-volume category, consistent with the broader prediction market industry. Presidential election contracts, congressional race outcomes, and major policy decisions attract both informed traders and casual participants. DraftKings’ sports betting user base skews toward the demographic that is also engaged with political wagering, creating natural liquidity.

Economic indicators represent a category where DraftKings Predictions competes directly with Kalshi, which has been listing economic contracts for years. Fed rate decisions, CPI prints, jobs reports, and GDP data are all binary-contract-friendly because they have unambiguous resolution criteria and official data sources. These markets tend to attract more sophisticated traders and see tighter spreads.

Entertainment and cultural events are a category where DraftKings Predictions can leverage its brand positioning in sports and pop culture. Award show outcomes, reality TV eliminations, and box office milestones appeal to the same audience that watches sports and makes prop bets. These markets often have lower liquidity but higher engagement from casual users.

Weather contracts are a newer category that taps into both genuine hedging demand (agricultural and energy sector participants who want to hedge weather risk) and speculative interest. Hurricane season activity, temperature records, and seasonal precipitation totals all have well-defined resolution criteria.

DraftKings Predictions vs. DraftKings Sportsbook

These are two products inside one app, and confusing them leads to fundamental misunderstandings about what you are building against. The following table captures every dimension that matters for developers and agents.

The most important row in this table for agent developers is “market mechanic.” On the sportsbook side, your agent is playing against DraftKings’ trading desk – a sophisticated operation with access to sharps consensus, line movement data, and real-time risk management. On the Predictions side, your agent is competing with a diverse set of market participants ranging from retail users who followed a social media tip to quantitative traders running models. The distribution of counterparty sophistication is much wider on the exchange, which creates more frequent opportunities for edge.

The shared wallet is the operational detail that matters most. A user (or agent) funded on DraftKings can deploy capital to either product without friction. For cross-platform arbitrage strategies, the ability to move funds between a sportsbook position and a prediction market position within the same platform eliminates an entire category of capital efficiency drag that exists when arbitraging across separate platforms.

Comparison: Polymarket vs. Kalshi vs. DraftKings Predictions

Three platforms now compete for event contract volume, each with fundamentally different architectures. Understanding where they differ is essential for any developer deciding where to build – or how to build across all three.

Where Polymarket Wins

Polymarket has three structural advantages that will persist. First, global access. Any trader anywhere in the world with a crypto wallet can trade on Polymarket, which means the platform aggregates global information into prices for events that are inherently international. Second, no position limits. CFTC-regulated exchanges must impose position limits on contracts; Polymarket does not, which means large traders (whales) can take concentrated positions that would not be possible on Kalshi or DraftKings Predictions. Third, the most mature API. Polymarket’s CLOB API has been iterated on for years and supports the full range of operations an automated trading agent needs.

Where Kalshi Wins

Kalshi’s advantage is regulatory clarity combined with API maturity. It is the longest-operating CFTC-regulated event contract exchange, with a documented API that supports REST, WebSocket, and FIX protocol access. For institutional participants and developers building compliant agents, Kalshi provides the clearest legal framework. Kalshi also wins on event breadth – it has listed more unique contract types across more categories than either competitor, including economic indicators where it has deep domain expertise.

Where DraftKings Predictions Wins

DraftKings Predictions’ advantage is distribution. No other event contract platform has access to millions of existing sports bettors who are already KYC-verified, funded, and accustomed to wagering on outcomes. This user base advantage means that as DraftKings Predictions matures, it has the potential to bring liquidity that neither Polymarket nor Kalshi can match from their native user bases alone.

The second advantage is the sportsbook-prediction market bridge. DraftKings is the only platform where a user can hold a traditional sports bet and an event contract in the same account, and where an agent could (once API access is available) monitor both product lines for correlated opportunities. This is not possible on Kalshi or Polymarket because they do not operate sportsbooks.

For a more granular platform comparison, see our dedicated breakdowns: DraftKings vs. Polymarket, Kalshi vs. DraftKings Predictions, and the three-way comparison.

The Arbitrage Opportunity

DraftKings Predictions creates arbitrage opportunities that did not exist before, precisely because it bridges two different market structures within the same company. Cross-product arbitrage between an event contract exchange and a sportsbook, operated by the same entity, is a new category of opportunity. Here is where the edges emerge.

DraftKings Predictions vs. DraftKings Sportsbook

This is the most novel arbitrage vector. The same company offers two products with different pricing mechanisms for events that sometimes overlap.

How it works: DraftKings Sportsbook might offer a prop bet on a major awards show at -180 (implied probability 64.3% before vig removal, approximately 60% true probability at standard vig). Simultaneously, DraftKings Predictions might price the equivalent Yes contract at $0.55. That is a 5-percentage-point gap. If your model estimates the true probability at 58%, you can buy the Yes contract on Predictions at $0.55 (expected value positive if true probability exceeds 55%) and potentially hedge with a No position or simply identify which side offers the better expected value.

Why the gap exists: The sportsbook price is set by DraftKings’ trading desk and adjusted based on handle distribution and risk management. The Predictions price is set by a market of individual traders. These two pricing mechanisms can diverge, especially on:

• Events where the sportsbook has limited data and sets wider vig to compensate for uncertainty
• Events where Predictions liquidity is thin and a few informed traders can move the price significantly
• Timing windows when new information hits one product before the other (a sportsbook line might react faster to breaking news than a thinly traded Predictions contract, or vice versa)
• Events that attract different participant types on each product (recreational bettors on the sportsbook vs. information traders on Predictions)

Operational advantage: Because both products share a single DraftKings wallet, you do not face the capital fragmentation problem that plagues cross-platform arbitrage. When you arbitrage Polymarket against Kalshi, your capital is split across two platforms with different funding mechanisms, settlement timelines, and withdrawal processes. When you arbitrage Predictions against Sportsbook, the capital is in one place.

DraftKings Predictions vs. Polymarket

This cross-platform arbitrage follows the same logic as any two-venue arbitrage: find events listed on both platforms, identify price discrepancies, and take opposite positions to lock in a profit.

Where discrepancies emerge: Political markets are the primary overlap zone. Both platforms list contracts on elections, policy decisions, and geopolitical events. Polymarket prices reflect a global, crypto-native trader base. DraftKings Predictions prices reflect a U.S. sports bettor base. These are different information pools, and they arrive at different probability estimates, particularly for events where U.S. domestic sentiment diverges from international consensus.

Practical challenges: The two platforms use different currencies (USDC vs. USD), different funding mechanisms, and different settlement processes. Capital is fragmented. An agent running this strategy needs to maintain funded accounts on both platforms and manage the currency conversion and withdrawal logistics. The cross-platform arbitrage guide covers the operational framework in detail.

DraftKings Predictions vs. Kalshi

Both platforms are CFTC-regulated DCMs listing binary event contracts, yet they serve different user bases and arrive at prices through different trader compositions.

The structural logic: Kalshi’s user base skews toward fintech-savvy retail traders and institutional participants. DraftKings Predictions’ user base skews toward sports bettors discovering event contracts for the first time. For the same economic indicator contract – say, “Will CPI exceed 3.0% for March 2026?” – these different trader populations may arrive at different probability estimates. An agent monitoring both order books can identify and exploit these discrepancies.

Regulatory note: Because both platforms are CFTC-regulated, an agent operating across both must be attentive to position limits, reporting requirements, and the possibility that correlated positions across venues might attract regulatory scrutiny. This is a more conservative arbitrage environment than the Polymarket-Kalshi pair, where one side operates outside U.S. regulation.

Building an Arbitrage Monitor

The practical starting point for any cross-DraftKings arbitrage strategy is a monitoring system that tracks prices across both DraftKings products and compares them against Kalshi and Polymarket. The architecture looks like this:

1. Data ingestion layer: Pull event contract prices from Kalshi (via REST API) and Polymarket (via CLOB API). For DraftKings Predictions and DraftKings Sportsbook, data collection currently requires web scraping or manual monitoring since no public API is available.

2. Event matching engine: Map events across platforms. “Will the Fed cut rates in June 2026?” might appear with different wording on each platform. The matching engine needs to normalize event descriptions and confirm that contracts across platforms have compatible resolution criteria.

3. Implied probability calculator: Convert sportsbook odds (with vig removal) and event contract prices into comparable implied probabilities. Apply fee adjustments for each platform to get net expected values.

4. Discrepancy detection: Flag pairs where the implied probability gap exceeds a configurable threshold after accounting for fees on both sides.

5. Execution layer: For Kalshi and Polymarket, API-driven execution is available today. For DraftKings Predictions, execution is manual until an API is released. Agents should be designed so that the execution layer is modular – adding a new venue is a configuration change, not a rewrite.

Agent Architecture for DraftKings Predictions

Building agents that can operate on DraftKings Predictions requires thinking about both the present constraints and the expected future state. The platform is not yet API-accessible for trading, but that does not mean developers should wait to build.

Current State of API Access

As of March 2026, DraftKings has not released a public trading API for DraftKings Predictions. The existing DraftKings API is oriented toward DFS contest data, sportsbook odds feeds, and affiliate/content integrations. It does not support event contract order submission, position management, or market data streaming for Predictions.

Internal endpoints exist – the DraftKings app communicates with backend services that expose event contract data, order book state, and trading functionality. These endpoints are undocumented, subject to change without notice, and using them programmatically would likely violate DraftKings’ terms of service.

What this means practically: you cannot build a fully autonomous trading agent on DraftKings Predictions today the way you can on Kalshi or Polymarket. But you can build everything around it.

What You Should Build Now

The correct approach is to build your agent architecture platform-agnostically, with DraftKings Predictions as a planned venue, and use Kalshi and Polymarket for live trading while the DraftKings API is unavailable.

Data monitoring agent. Build an agent that monitors DraftKings Predictions market data – contract prices, volume, event listings, and order book depth – even if it cannot trade. This data is visible through the DraftKings app and website. An agent that has been collecting DraftKings Predictions price history for months will have a significant advantage when the API launches, because it will already have training data for models, baseline understanding of market microstructure, and calibrated expectations for liquidity across event categories.

Cross-platform signal agent. Build an agent that uses DraftKings Predictions prices as a signal source even if it cannot execute on the platform. If DraftKings Predictions prices a political event at $0.70 Yes while Kalshi prices it at $0.64, the DraftKings price provides signal about how the sports-bettor population views the event. Your agent can trade on Kalshi (where it has API access) using DraftKings Predictions as an additional information source.

Abstracted order management. Design your order management system with a platform-agnostic interface. Each venue (Kalshi, Polymarket, DraftKings Predictions, future venues) should implement the same interface: submit order, cancel order, get positions, get market data. The Kalshi and Polymarket implementations are live today. The DraftKings implementation is a stub that logs intended actions and waits for the API. When the API ships, you fill in the implementation without touching any other code.

// Platform-agnostic interface design
interface EventContractVenue {
  getMarkets(): Market[]
  getOrderBook(marketId: string): OrderBook
  submitOrder(order: Order): OrderConfirmation
  cancelOrder(orderId: string): void
  getPositions(): Position[]
  getBalance(): Balance
}

// Implementations
class KalshiVenue implements EventContractVenue { /* live */ }
class PolymarketVenue implements EventContractVenue { /* live */ }
class DraftKingsPredictionsVenue implements EventContractVenue { /* stub */ }

This is the architecture pattern described in our prediction market API reference – build the abstraction today so that adding venues tomorrow is incremental.

Cross-Platform Agent Design

The highest-value agent architecture in the converged landscape is one that operates across multiple event contract platforms and sportsbooks simultaneously. Here is the reference architecture.

Layer 1 – Data aggregation. A unified data pipeline that ingests market data from every connected venue. For prediction markets, this means order book snapshots and trade data via API. For sportsbooks, this means odds feeds via aggregators like The Odds API or direct scraping. For DraftKings specifically, this layer needs to handle both Predictions data and Sportsbook odds data, keeping them distinct but correlated.

Layer 2 – Event matching and normalization. A service that maps events across platforms. The same real-world event appears with different identifiers, different wording, and sometimes different resolution criteria on each venue. The matching service must handle fuzzy matching, resolution criteria comparison, and temporal alignment (different platforms may have different contract expiration times for the same underlying event).

Layer 3 – Pricing and signal generation. Models that consume normalized data from all venues and produce probability estimates. These models can range from simple cross-platform averaging (the consensus probability is the volume-weighted average of all venue prices) to sophisticated Bayesian models that weight each venue’s price based on historical accuracy and participant composition.

Layer 4 – Strategy and execution. Decision engines that identify trades based on the pricing layer’s output. Strategies include:

• Pure arbitrage (simultaneous opposite positions on two venues for the same event)
• Cross-product arbitrage (sportsbook vs. prediction market for correlated events)
• Market-making (providing liquidity on thin markets and earning the spread)
• Informed trading (using model estimates to take directional positions where the market price deviates from the model)

Layer 5 – Risk management. Portfolio-level risk management across all venues. This layer tracks aggregate exposure by event, manages position limits per venue (especially important for CFTC-regulated platforms), monitors capital allocation across venues, and enforces drawdown limits.

For a deeper dive into agent stack components, see the agent betting stack guide.

Risk Management Considerations

DraftKings Predictions introduces risk management considerations that pure prediction market agents do not face.

Regulatory risk. DraftKings Predictions operates under CFTC regulation through its DCM designation. Position limits apply. An agent that also trades on Kalshi (another DCM) must consider whether correlated positions across two CFTC-regulated venues create aggregate exposure that might trigger reporting thresholds or attract scrutiny.

Platform risk. DraftKings can change Predictions’ rules, fee structure, or available markets at any time. The product is still maturing, and early-stage event contract platforms frequently adjust their parameters. An agent dependent on specific fee assumptions or market availability should build those assumptions as configurable parameters, not hardcoded values.

Liquidity risk. DraftKings Predictions liquidity is still growing. Thin markets mean wide spreads, slippage on larger orders, and the possibility that exit liquidity will not be available when you need it. Agents should size positions based on observed order book depth, not on the notional contract value.

Counterparty risk. On DraftKings Predictions, your counterparty on settlement is DraftKings itself (as the exchange operator). DraftKings is a publicly traded company with substantial financial resources, making counterparty risk low. This is a meaningful advantage over Polymarket, where settlement depends on on-chain infrastructure and the UMA oracle system.

What This Means for the Industry

DraftKings Predictions is not just a new product feature. It represents a structural shift in how the wagering industry operates, and the implications ripple through every layer of the ecosystem.

The Sportsbook-Exchange Hybrid Is Here

DraftKings is now simultaneously a bookmaker and an exchange operator. It sets odds and takes the other side of sports bets on its sportsbook, and it operates a neutral matching platform where users trade event contracts against each other on Predictions. These two models have fundamentally different economics: the bookmaker profits from the vig and loses when bettors win, while the exchange profits from fees regardless of outcomes.

No other company in the U.S. occupies this dual position at scale. Kalshi is an exchange but not a sportsbook. FanDuel is a sportsbook but has not launched a DCM-regulated prediction market. Polymarket is a crypto-native exchange with no sportsbook operations. DraftKings is alone in straddling both models, and this gives it unique strategic options – cross-selling users between products, sharing data and risk models across the two businesses, and offering a single-wallet experience that lowers friction for users who want to participate in both.

For developers building AI agents that operate across both paradigms, DraftKings is the first platform where this cross-paradigm operation happens within a single account. That changes the math on capital efficiency, the logistics of cross-product arbitrage, and the data richness available to agents that can observe both sportsbook odds and prediction market prices simultaneously.

Regulatory Convergence Accelerates

DraftKings Predictions operating under a CFTC DCM license while DraftKings Sportsbook operates under state gaming licenses creates an interesting regulatory dynamic within a single company. The same user, in the same app, is subject to two different regulatory frameworks depending on which tab they are on.

This dual-regulation model will pressure regulators to clarify boundaries. When does an event contract become a sports bet? When does a sports prop become a binary contract? The Kalshi lawsuit over sports event contracts opened this question at the federal level, and DraftKings Predictions operating alongside DraftKings Sportsbook makes the question practically urgent.

For developers, the near-term implication is that you should build compliance awareness into your agent architecture at the venue level. Each venue has its own regulatory constraints – position limits, reporting requirements, permitted event types, geographic restrictions. An agent that treats all venues identically will eventually run into compliance issues. An agent that maintains venue-specific compliance profiles can adapt as the regulatory landscape evolves.

The API Race

DraftKings Predictions’ lack of a public API is its most significant limitation for the developer community, and it is almost certainly temporary. Here is why:

Market-maker demand. Thin order books hurt the platform. Market makers who provide liquidity and tighten spreads need API access to manage their positions programmatically. Every event contract exchange that has achieved meaningful liquidity has done so in part by onboarding automated market makers through API programs. DraftKings will face the same imperative.

Competitive pressure. Kalshi and Polymarket both offer comprehensive APIs that support fully automated trading. If DraftKings Predictions wants to attract the quantitative trading community – which provides the lion’s share of liquidity on event contract platforms – it needs to offer comparable programmatic access.

Developer ecosystem. DraftKings has historically positioned itself as developer-friendly through its DFS and sportsbook data APIs. Extending this positioning to Predictions is a natural step, especially as the company competes with Kalshi and Polymarket for the attention of developers building prediction market agents.

When the API does launch, developers who have already built platform-agnostic agent architectures with DraftKings Predictions as a planned venue will be in the strongest position to move quickly. The pattern is the same one that played out when Kalshi expanded its API capabilities and when Polymarket launched its CLOB – early adopters who had built flexible architectures captured the most value.

Distribution Wins

The prediction market industry has a distribution problem. Polymarket has deep liquidity but its user base is overwhelmingly crypto-native. Kalshi has regulatory clarity but its user base is niche. Neither platform has cracked mainstream consumer adoption at the scale that sportsbooks have.

DraftKings Predictions bypasses the distribution problem entirely. It does not need to convince new users to create accounts, go through KYC, and fund wallets on an unfamiliar platform. Millions of DraftKings users are already there, already verified, already funded. The conversion from “sportsbook user” to “event contract trader” is a tap within the app, not a new account signup.

This distribution advantage means that DraftKings Predictions has the potential to onboard more event contract traders in a year than Kalshi or Polymarket have in their entire histories. If even a small fraction of DraftKings’ sportsbook user base engages with Predictions, the resulting liquidity could transform the event contract landscape.

For the industry as a whole, this is net positive. More traders on event contracts means tighter spreads, more accurate prices, and more opportunities for agents that provide liquidity or trade on information. It also means that the tools, guides, and infrastructure the developer community builds for event contract trading will have a larger addressable market – which is precisely why this article exists on AgentBets.

Frequently Asked Questions

What are DraftKings Predictions?

DraftKings Predictions is a CFTC-regulated event contracts platform that lets users trade binary outcome contracts on political, economic, and other events. It operates separately from DraftKings Sportsbook, regulated under federal CFTC oversight rather than state gaming commissions. It was built through DraftKings’ acquisition of Railbird, a CFTC-registered designated contract market (DCM).

How is DraftKings Predictions different from DraftKings Sportsbook?

DraftKings Sportsbook is a traditional sportsbook regulated by state gaming commissions where you bet against the house at odds DraftKings sets. DraftKings Predictions is an exchange-based prediction market regulated by the CFTC where users trade binary event contracts with each other. Different regulation, different mechanics, different markets. The sportsbook is a dealer market; Predictions is an exchange. For a complete breakdown, see the comparison table above and our guide on sports betting vs. prediction markets.

Can you arbitrage between DraftKings Predictions and DraftKings Sportsbook?

Yes, there are arbitrage opportunities when DraftKings Predictions event contract prices diverge from implied probabilities on DraftKings Sportsbook for the same or correlated events. The shared wallet makes capital management easier than cross-platform arbitrage. See the arbitrage section above for detailed strategies and our cross-platform arbitrage guide for the operational framework.

How does DraftKings Predictions compare to Polymarket?

DraftKings Predictions is CFTC-regulated, requires full KYC, is US-only, uses USD, and is backed by a publicly traded company. Polymarket is crypto-based, has minimal KYC for small trades, is globally accessible, and operates outside traditional regulatory frameworks. Polymarket currently has significantly higher liquidity and volume on most markets, plus a mature public API. DraftKings has the advantage of a massive existing user base. See the three-way comparison table above and our dedicated DraftKings vs. Polymarket comparison.

How does DraftKings Predictions compare to Kalshi?

Both are CFTC-regulated DCMs offering binary event contracts to US users with full KYC. Kalshi was first-to-market, has broader event categories, and offers a mature trading API with REST, WebSocket, and FIX protocol access. DraftKings Predictions benefits from DraftKings’ massive user base and brand recognition. See our Kalshi vs. DraftKings Predictions comparison for the full breakdown.

Is there an API for DraftKings Predictions?

As of early 2026, DraftKings has not released a full public trading API for DraftKings Predictions comparable to Kalshi’s REST/FIX API or Polymarket’s CLOB client. Internal endpoints exist but are undocumented and using them would likely violate terms of service. Given DraftKings’ developer-friendly positioning and the competitive pressure from Kalshi and Polymarket, a public API is anticipated. See the agent architecture section above for how to build now in preparation.

Should I build agents for DraftKings Predictions now or wait for the API?

Build now, but build platform-agnostically. Design your agent architecture with a venue abstraction layer so that adding DraftKings Predictions is a configuration change when the API launches. In the meantime, use Kalshi and Polymarket for live trading and monitor DraftKings Predictions data as a signal source. See our prediction market API reference for the abstraction pattern.

Can AI agents legally trade on DraftKings Predictions?

CFTC-regulated exchanges generally permit automated and algorithmic trading, subject to platform-specific rules and position limits. DraftKings Predictions’ terms of service will govern what is permitted once programmatic access is available. For a broader analysis of agent legality across platforms, see our AI sports betting agents guide.

What happens if the same event is priced differently on DraftKings Predictions and DraftKings Sportsbook?

This is a cross-product arbitrage opportunity. The sportsbook price includes embedded vig and is set by DraftKings’ trading desk. The Predictions price is set by market participants on an order book. When these prices diverge, the difference represents either a mispricing opportunity or a reflection of different information being available to different participant pools. An agent monitoring both can identify and potentially exploit these gaps.

Is DraftKings Predictions available in all states?

No. DraftKings Predictions’ availability depends on both federal CFTC regulation and individual state policies regarding event contracts. Some states have opted out of permitting event contract trading, and DraftKings may roll out the product incrementally across states. Check DraftKings’ current availability map for the latest state-by-state status.