There is a standard narrative in the sportsbook industry about prediction markets: they are a niche product, primarily attracting politically-minded retail traders and quant bettors who were never valuable sports wagering customers anyway. DraftKings CEO Jason Robins gave the most visible version of this argument in early 2026, dismissing migrating users as “low-margin or negative-margin.”
The data tells a different story. Prediction markets have grown at a rate that has no precedent in U.S. regulated gambling history. The users moving to platforms like Kalshi overlap directly with active sportsbook customers. And the behavioral signals that precede migration—bet sizing compression, market type drift, event-window drop-off—are detectable weeks before a user’s wallet actually shifts. The problem is that most CRM infrastructure is not looking for them.
The SurgeFrom 600K to 5.1M: How Fast Prediction Markets Actually Grew
Kalshi’s monthly active user base grew from approximately 600,000 at the start of 2025 to 5.1 million by year end—a ~750% increase in a single calendar year. Weekly trading volume increased 1,000% year-over-year to more than $1 billion per week. During a single NFL weekend, Kalshi processed $720 million in handle, including over $100 million on one game (Yahoo Finance, 2026).
These are not incremental metrics. They represent the fastest verified user acquisition trajectory in U.S. regulated gambling—faster than the DFS boom, faster than the post-PASPA sportsbook expansion in any single state, faster than any mobile gambling product launch on record. The scale matters because it establishes that this is not a niche migration. It is a structural shift in where a portion of the sports wagering population is choosing to place bets.
Industry acknowledgment arrived quickly. Six major sportsbook and fantasy operators launched competing prediction market products during 2025 and 2026. Their products gained almost no traction: Kalshi received 19 times more app downloads than DraftKings and FanDuel’s combined prediction market apps in January 2026 alone—1.9 million versus fewer than 100,000 combined (Fortune, February 2026). Native PM platforms dominate cross-product user acquisition. Sportsbook-branded PM products are not substitutes.
10% of DraftKings Users Are Already on Kalshi—and Growing
The most operationally significant data point in this analysis is not the headline growth number. It is the overlap. According to Apptopia cross-app usage data, approximately 10% of DraftKings’ user base was simultaneously active on Kalshi in January 2026. That overlap had grown every month since August 2025 (Fortune, February 2026).
This is not a stable dual-use equilibrium. It is an accelerating trend. Month-over-month growth in the overlap means that the share of DraftKings customers who are also testing prediction markets is increasing, not plateauing. For CRM purposes, this distinction matters enormously: a plateau suggests dual-platform loyalty; an accelerating overlap suggests an active substitution process that has not yet completed.
The geographic dimension adds another layer operators are underweighting. Prediction markets operate under CFTC jurisdiction, making them available in all 50 states. California, Texas, and a number of other large-population states where sportsbooks cannot legally operate are fully accessible PM markets. Operators who have been physically barred from these user segments now face a competitor with no such restrictions. The addressable market for prediction markets is structurally larger than for any individual sportsbook.
DraftKings’ framing of migrating users as low-margin is worth examining analytically. Apptopia’s cross-app data shows these users are actively engaging both platforms simultaneously—which is not the behavior of someone who was never generating sportsbook value. It is the behavior of someone currently generating sportsbook value who is also testing an alternative. The CRM risk is almost certainly being systematically underestimated.
Wallet BehaviorSubstitution-Plus-Expansion: The Signal CRM Tools Are Not Reading
The wallet-level data is where the CRM implications become concrete. After a user’s first prediction market bet, core sportsbook wallet drops 11%—but total gambling wallet expands 9% (Covers, February 2026). This is partial substitution without full churn. The user is not leaving gambling; they are reallocating gambling budget toward prediction markets while maintaining some sportsbook activity.
The pattern creates a narrow, measurable intervention window. After the first PM bet, the -11% wallet shift is not immediate full substitution—but it is the beginning of a reallocation process. Left unaddressed, partial substitution hardens into habit, and habit hardens into platform loyalty elsewhere. The users who experience a -11% shift and receive no CRM response from their sportsbook are the ones who show up in the cross-app data six months later as fully migrated.
The product category distinction operators cite—“prediction markets are different from sports betting”—does not hold up under volume analysis. During 2025, 87% of Kalshi’s $24 billion total volume came from sports event contracts. During football season, that share rose to 90% (Next Event Horizon, 2025). Only roughly $3 billion of Kalshi’s 2025 volume was non-sports. These are sportsbook customers betting on sports outcomes through a different regulatory vehicle—not a different product category.
The segment most disproportionately migrating is analytically sophisticated bettors—often called “sharps”—who are routinely limited or excluded by sportsbook risk controls. These users find prediction market liquidity more accommodating. From a sportsbook perspective, sharps are often characterized as low-margin because they win more frequently. But they also represent high engagement, high betslip volume, and high referral value. Their migration represents a CRM loss that is not visible in standard churn metrics precisely because sportsbooks may already be restricting their wager amounts.
Competitive ThreatKalshi at 20% of DraftKings Revenue in Under 12 Months
Kalshi’s annualized sports revenue reached $1.3 billion—already approximately 20% of DraftKings’ estimated $6.7 billion 2026 revenue. That competitive position materialized in under 12 months (Covers, February 2026). The canonical framework for competitive threat analysis—market share percentage, time to reach threshold, addressable user overlap—all point in the same direction.
Super Bowl LX is the sharpest single-event illustration. Prediction markets captured an estimated $630 million in bets on that game—representing roughly 80% of all year-over-year handle growth for the event (Fortune, February 2026). Traditional sportsbooks saw flat to declining handle on the largest single wagering event in the U.S. calendar. Analyst response was immediate: Flutter EPS estimates were cut 49% and DraftKings down 29% over three months, with downgrades directly attributed to prediction market cannibalization.
A broader estimate corroborates the single-event data: prediction market growth is correlated with an approximately 5% decline in legal sportsbook handle since Kalshi expanded to all 50 states (Covers, February 2026). A 5% handle decline at the industry level is not noise. It is the aggregate expression of the wallet-level substitution pattern described above, multiplied across millions of users.
Event-Driven WindowsMarch Madness: $4B Handle and 12% Dormant Reactivation
Major sporting events concentrate CRM leverage in ways that are especially relevant to the prediction market threat. March Madness 2026 is projected to generate $4 billion in sportsbook handle—a 6.7% increase from 2025’s $3.7 billion—plus an estimated $530 million in prediction market equivalent handle, for a combined $4.5 billion wagering event (World Casino Directory, 2026). Gross gaming revenue is projected at $279 million, a 23% year-over-year increase, as hold rate improves from 6.1% to 7%.
The dormant reactivation data embedded in March Madness is particularly actionable. Approximately 12% of March Madness 2026 bettors placed zero wagers before the tournament window. These are users who were acquired, sat dormant, and reactivated specifically because of a high-profile sporting event. The median handle per reactivated user during the tournament window was $5,127—a figure that makes the CRM value of catching these users before prediction markets do measurable and substantial.
The event-window pattern generalizes: NFL playoffs, Champions League knockout rounds, and World Cup cycles all create discrete moments where dormant sports bettors re-engage with wagering. If prediction market platforms are positioned for those moments with superior mobile UX, 50-state availability, and no deposit friction, the reactivation that operators should be capturing flows to Kalshi instead. BidCanvas research on dormant reactivation covers the mechanics of event-window CRM intervention in detail.
CRM Blind SpotWhy Current CRM Infrastructure Misses These Signals
The structural problem with sportsbook CRM tools is where they look for churn. Standard churn models flag users after betting frequency drops or deposit activity ceases. Cross-platform migration shows no in-platform signal until the process is largely complete. A user who has shifted 30% of their gambling wallet to Kalshi is still placing sportsbook bets at reduced frequency and stake—every individual in-platform metric still looks like light activity rather than migration.
The -11% wallet shift happens gradually, over weeks. Without behavioral fingerprinting at the betslip level—changes in stake sizing, shifts in market type preferences, changes in event-window participation patterns—operators cannot distinguish between a user who is having a slow month and a user who is actively reallocating budget to a competitor. The behavioral signatures are different, but only visible in the granular data.
DraftKings’ public dismissal of migrating users as low-margin may reflect genuinely different segment economics for their specific user mix. Or it may be post-hoc rationalization of a loss the CRM tooling cannot detect. The Apptopia data—showing month-over-month overlap growth with an accelerating trend—is inconsistent with the claim that these users were not generating sportsbook value. Accelerating overlap is the behavioral signature of active migration from a platform you are still using, not abandonment by users who were never engaged.
Operators with betslip-level data—stake patterns, market preferences, odds sensitivity, event coverage breadth—have the inputs to build early-warning cross-product migration models. The data is already being collected. The analytical layer to interpret it as migration risk rather than normal variance is what most CRM stacks currently lack. This is where AI Betslips operates: instrumenting the betslip-level behavioral signals that precede wallet shift.
What To Do NowThe Three Signals Operators Should Be Tracking Today
The research data points toward three specific behavioral signals that are detectable at the betslip level before wallet substitution becomes visible in aggregate CRM metrics.
Signal 1: Bet Sizing Compression
Users who begin reducing average stake by more than 15% over a rolling 30-day window while maintaining bet frequency are often in the early stages of testing PM liquidity. The pattern is consistent with a user running parallel experiments: placing familiar sportsbook bets at reduced size while allocating the freed capital to prediction market contracts. Frequency maintenance is the key distinguishing feature—this is not casual disengagement. It is deliberate reallocation.
Signal 2: Market Type Shift
Migration away from spread and totals markets toward moneylines mirrors the contract structure of prediction markets. Prediction market contracts are binary outcome instruments—structurally identical to moneyline bets. A user shifting their market type mix toward moneylines, particularly on events where PM alternatives exist, is displaying a behavioral preference that correlates with PM testing. This signal precedes the stake compression in many migration patterns, making it an earlier-stage indicator.
Signal 3: Event-Window Drop-Off
Users who skip high-volume event windows—NFL playoffs, March Madness, Super Bowl—without a traceable churn reason have likely found PM alternatives for those events. The 12% of March Madness 2026 bettors who placed zero pre-tournament wagers and then became high-value event-window bettors demonstrates the reactivation value at stake. The same logic applies in reverse: a user who was active in previous event windows and misses the next one is not dormant—they may be placing the same bets on Kalshi instead.
The practical implementation path runs through betslip-level data instrumentation. BidCanvas AI Betslips captures stake patterns, market-type drift, and event-window behavior at the individual betslip level—the granularity required to distinguish migration signals from normal behavioral variance. When the signal pattern fires, the CRM integration triggers a retention campaign targeted at the specific behavioral fingerprint, not a generic churn risk flag.
The prediction market growth story will not pause while operators build detection infrastructure. Kalshi’s 750% MAU growth and the accelerating DraftKings overlap are both trailing indicators of a migration process that started months before those numbers appeared in reporting. The leading indicators are in the betslip data, available today, in every operator’s existing data warehouse. The question is whether the analytical layer exists to read them as migration risk before the wallet has already shifted.
For operators thinking through how prediction markets affect CRM strategy more broadly, or modeling the dormant reactivation opportunity that event windows create, the underlying behavioral signal infrastructure is the same: betslip-level granularity, interpreted against a cross-product behavioral model, delivered to CRM tooling as an actionable segment. The technology exists. The data exists. The migration is happening now.
SourcesData Sources & Attribution
- Yahoo Finance — Prediction Markets Surge, Eclipse DraftKings — Kalshi MAU growth (600K → 5.1M), weekly volume 1,000% YoY increase
- Fortune — Gambling Stocks, DraftKings, Flutter, Prediction Markets (February 2026) — 10% cross-app overlap, app download data, Super Bowl LX PM handle, EPS estimate cuts
- Covers — Kalshi Prediction Markets Stealing Market Space (February 2026) — -11%/+9% wallet shift, $1.3B annualized sports revenue, ~5% sportsbook handle decline
- Next Event Horizon — Prediction Markets Are Mostly Sports Betting — 87% sports share of Kalshi 2025 volume
- AIInvest — Rise of Prediction Markets (2025) — $27.9B total PM volume, 47% annual growth, $100.9B global sports betting market context
- World Casino Directory — March Madness Betting Expected to Reach $4B in 2026 — $4B projected handle, $279M GGR, 12% dormant reactivation cohort, $5,127 median handle