Prediction Markets & the Financialization of Gambling
A PayPal Ventures Sensing Report
December 2025

Rachel Zabronsky & Amman Bhasin

The opinions expressed in this blog are solely the author’s and do not reflect the views of PayPal.

Everything Is Financialized

In a bygone era, if you wanted to gamble you went to Vegas. Today, the friction that once kept gambling attached to casinos, racetracks, and local bookies has vanished: it is bigger, pervasive, faster, and almost entirely mobile. Sports betting is now a mainstream consumer product, reaching $150 billion wagered in the U.S. alone in 2024 – a number that is only growing (ESPN). Just as importantly, the cultural stigma around gambling has largely faded. Gambling-related information like in-game odds is featured in sports broadcasts and built into the same digital flows as shopping or social media.

The lines between investing and gambling are becoming increasingly blurred. People flip sneakers, meme coins, Pokémon cards, and Sunday football parlays with the same instincts that drive stock trades on Robinhood. The act of placing a bet and the act of making a trade now look almost identical.

Prediction markets have emerged inside this same cultural moment. In short, prediction markets are exchanges that let you trade on the outcome of events: from sports to elections to inflation to Rotten Tomatoes scores. They turn every question – will Jaxson Dart throw a touchdown pass? Who will win the election? Will inflation fall below 3%? – into a market-driven asset.

At PayPal Ventures, we’re interested in this shift because it sits at the intersection of financial infrastructure, regulation, and consumer behavior. The mechanics of gambling are merging with the tools of finance, and the resulting platforms are shaping the next generation of participatory markets.

Here, we’ll explore the rise of prediction markets and what they might mean for the landscape at large. The authors (Amman and Rachel) come at this from opposite sides of the table: one of us has a soft spot for a good parlay, the other would rather watch from the sidelines. While researching this piece, one of us was betting on what Jerome Powell would say next, and the other was more interested in wagering on Rotten Tomatoes scores and the next Taylor Swift album release. We’ll let you guess who’s who. We are also having healthy debates on several of the topics we discuss below, such as how the regulatory landscape might evolve, whether prediction markets become a winner-take-all market, and if non-sports categories can really scale – but that’s exactly why we find the topic so interesting.

From Sportsbooks to Marketplaces

Sports betting has exploded in the U.S. since 2018, when the Supreme Court opened the door for state-by-state legalization (NYT). DraftKings, FanDuel, and others turned watching games into an interactive experience where fans could back their instincts with real money. Betting became entertainment, and the entertainment became commerce.

But traditional sportsbooks still operate as the house. They set the odds, take the other side of each bet, and hold a built-in margin (“the vig”) on every outcome. The bettor is effectively a customer buying a product, not a trader making a market. Only a small fraction of users actually win, and those who do rarely last long before limits kick in.

Prediction markets flip that model. They behave like exchanges, with users buying and selling positions in the outcome of an event instead of betting against a sportsbook. Each share trades at a price that reflects the collective market probability. For example, if a “Yes” share on “the Golden State Warriors make the playoffs” is priced at 70 cents, it implies a 70 percent likelihood. If they make it, the share settles at a dollar; if not, it goes to zero.

This system allows for continuous price discovery as new information enters the market. The probability estimates (and value) associated with outcomes move the same way a stock price reacts to earnings or news. The structure attracts both retail users who want to speculate as well as institutional market makers, creating a liquid, balanced, efficient ecosystem that can feel closer to finance than gambling.

Prediction markets reach far beyond sports. They host trading on elections, economic data, policy outcomes, and cultural moments. A user can just as easily take a position on inflation trends, if Paul Thomas Anderson will win Best Director, or what the weather in San Francisco will be tomorrow. Together, these markets form a live snapshot of collective expectations – a dynamic, market-based view of what people believe will happen next.

These platforms now look more like exchanges than casinos, with transparent order books, market makers, and APIs that allow others to build on top. Prediction markets are transforming betting into a liquid, data-driven, and socially powered form of trading.

The Regulatory Split: State vs. Federal

The U.S. is running two parallel experiments. On one track, gambling is regulated on a state-by-state basis. Every state has its own licensing rules, tax regimes, and approaches to enforcement. That model made sense when gambling was anchored to physical casinos or local sportsbooks, but it starts to look outdated once markets move online and money flows digitally and across borders. 30 states have legalized online sportsbooks (NYT), which leaves a huge swath of Americans who do not have legal access to online sportsbooks like DraftKings or FanDuel.

The other track involves federal oversight under the Commodity Futures Trading Commission (CFTC). The idea is that event contracts – markets tied to real-world outcomes – preempt state gambling laws. Platforms like Kalshi and Robinhood argue these contracts should be treated as financial derivatives rather than bets, falling exclusively under federal commodities law. Their legal wins around election markets in 2024 represented a potential turning point: if courts affirm that CFTC-approved event contracts supersede state gambling restrictions, it could set a new rulebook for everything that follows (CNBC).

If event contracts are accepted as derivatives, they must comply with the suite of CFTC regulation: DCMs (Designated Contract Market), DCOs (Derivatives Clearing Organization), clearing, margin, registration, surveillance, reporting, etc. That is a much higher operational bar, but the upside is national reach, consistent rules, and less state-by-state negotiation. The licenses required to operate a prediction market with federal regulation under the CFTC are extremely hard to come by; there are just over 20 active DCM and DCO licenses in the U.S. today (CFTC). Companies looking to get into the prediction market space are acquiring licenses through acquisitions instead of going through the approval process on their own: Polymarket, for example, purchased the regulated exchange QCEX in July 2025 (Financial Times); DraftKings acquired Railbird in October 2025 (DraftKings); and Robinhood acquired LedgerX, a CFTC licensed DCM, DCO, and Swap Execution Facility, in November 2025 (Reuters).

Some prediction markets began under sweepstakes models, especially in states with restrictive gambling laws. Sweepstakes operations argue they’re not “gambling” because there’s no direct payment for bets. Users earn “virtual credits” via referrals or tasks and then use those credits to play. That structure has come under increasing pressure. In the first half of 2025, states including Connecticut, Maryland, and Michigan issued cease-and-desist orders, passed new statutes, or filed lawsuits against sweepstakes operators (American Gaming Association). The sweepstakes window, once a regulatory workaround, is shrinking.

There are several approaches to operating nationally – here’s what some of the biggest and most interesting players are doing:

• Kalshi: holder of both a DCM and a DCO license, they are fully embedded in the derivatives path and hold a CFTC registration.
• Polymarket: runs like a DeFi exchange (fully transparent, borderless, and fast) but is not currently legally accessible in the U.S. They recently made an acquisition (QCEX) which will give them access to DCO and DCM licenses, and received formal approval from the CFTC to resume U.S. operations, allowing them to resume U.S. operations, as of December 3, 2025 (X,Reuters).
• Robinhood: doesn’t own a full prediction exchange itself but leverages Kalshi’s license to offer event contract markets to its user base. That setup allows Robinhood to offer prediction markets through its partnership with Kalshi without directly holding the regulatory burden. Robinhood has agreed to acquire MiamiX, potentially as an attempt to bring prediction market capabilities in-house.
• Novig: a sports-focused prediction market upstart which is currently operating under the sweepstakes model but exploring obtaining CFTC certification (CDC Gaming).
• Sweepstakes / DFS operators (e.g., PrizePicks): seeking regulatory legitimacy by acquiring derivative-related licenses (like Futures Commission Merchant (FCM) registrations), having them interface with true event contract platforms under regulated stacks.
• Coinbase & Gemini: Coinbase hinted at its intention to launch a prediction market product in addition to other new products like tokenized equities (Yahoo Finance). Gemini is also reportedly working on launching a prediction market product and filed a DCM application with the CFTC (Bloomberg).
• Sports books: DraftKings (via a partnership with Polymarket and its acquisition of Railbird) and FanDuel (via partnership with CME, a global derivatives marketplace) both appear to be gearing up to enter the space.

What people often miss is how much is at stake. States currently collect billions in taxes from legal sports betting. In 2024, licensed sportsbooks across the U.S. handled nearly $150 billion in wagers and generated over $14.2 billion in operator revenue; the state tax take was about $2.9 billion (Legal Sports Report). If prediction markets are subject to federal regulation, those state tax flows could be significantly reduced, evolve into new types of revenue, or disappear altogether.

At the same time, the federal tax code quietly tilts the playing field: gambling losses are deductible only up to winnings, and even then, only 90% under current limits, due to the language in The One Big Beautiful Bill Act. That asymmetry creates another incentive for users and operators to favor CFTC-licensed event-contract exchanges, where trades may qualify as financial instruments with full loss recognition.

On the regulatory front, major players are increasingly under scrutiny, and we expect this will only continue. Below are some of the most important and recent lawsuits we’re seeing play out in real time:

• State & Tribal Nation Opposition: The conflict is marked by escalating action from state regulators and tribal nations who classify sports-related prediction markets as illegal gambling. Beginning with cease-and-desist orders from New Jersey and Nevada in August 2025, a growing list of states (including Massachusetts, New York, and Connecticut) have filed lawsuits, while several California tribal nations have filed separate suits alleging violations of Indian gaming lawsThe Federal Preemption Defense: Prediction market operators like Kalshi and Robinhood are fighting back by filing federal lawsuits, asserting that their event contracts are federally regulated derivatives under the CFTC's exclusive jurisdiction, which they claim should legally preempt all conflicting state gambling laws..
• Despite this turmoil, prediction markets have real traction. Kalshi is on track to do $50 billion in trading volume on an annualized basis, a big jump from the roughly $300 million it recorded last year (NYT). In October, Kalshi recorded $4.4B in monthly trading volume, while Polymarket recorded $3B (The Block). In September, Robinhood’s CEO tweeted that total lifetime event contracts crossed 4 billion, with over 2 billion in a single quarter (X).

All of this points toward a larger fight: who regulates these products? States fear a loss of control and tax revenue. Operators push for predictability, scale, and national access. The courts are going to be the battleground for that question. If the federal path wins, we may see prediction markets operate nationwide, but states that have yet to legalize sports betting may resist on moral grounds, and those that allow it will likely resist giving up revenue.

Models & Product Shapes in Prediction Markets

The prediction market space is already splitting into distinct archetypal and product camps. What sets each model apart is not just branding, but how they generate revenue, manage liquidity, and scale over time. Below are two core structural archetypes and the product tensions they face.

The models themselves are not static. Platforms may begin with white-label rails, then build outward, or start consumer-first but later spin off a B2B arm.

Structural Models

White-label infrastructure / B2B rails
When a major brand (Robinhood, Fanatics, a major sportsbook) embeds prediction mechanics into their core product, user friction drops. Robinhood’s integration with Kalshi, for example, lets their 25-30 million users trade event contracts (“prediction markets”) directly through their app. That kind of embedding forces a shift from a siloed betting product to a feature in a broader financial or entertainment stack.

In this model, the prediction market operator acts more like a clearing layer or exchange than a consumer brand. They power markets behind the scenes, letting brokers, apps, or platforms plug in. Kalshi is a prime example: they already power event contracts through Robinhood via their derivatives arm. (Robinhood)

The benefit here is scale without marketing risk. The exchange concentrates liquidity and compliance, while front-ends focus on user experience. And as Robinhood grows, it may even seek its own license to re-bundle prediction services natively, displacing Kalshi.

DraftKings, Fanatics, and FanDuel are all rumored to be actively exploring entry into the space as well.

Standalone consumer platforms
These are consumer-facing brands that own everything: product, liquidity, marketing, and brand. Polymarket and Novig fit this mold (while Kalshi straddles both lines: it’s powering external apps while also maintaining its own front-end presence).

This is a riskier approach: you need to build brand, trust, liquidity, and network effects from scratch. But if these companies succeed, they control the user relationship end-to-end, which is powerful for monetization, engagement, and ecosystem control.

As we explore the prediction markets space from a venture capital perspective, the cold-start problem is a challenge new startups in this category will have to overcome. Building a standalone consumer platform against enormous incumbents like Robinhood, Kalshi, and Polymarket requires a significant amount of capital and time to build brand reputation.

Product & Monetization Differences

Even when platforms share the same regulatory framework, they can diverge sharply in how they make money, how “sticky” users are, what features they offer, and how they manage risk. Below are a few axes of differentiation:

Fee structures & liquidity incentives
Prediction markets make money differently than sportsbooks. Instead of taking the other side of a bet and building in a house edge, exchanges generate revenue by facilitating trades.

Kalshi operates more like a traditional exchange. It charges transaction fees on each trade, taking a small percentage of the expected earnings from both buyers and sellers. The platform doesn’t bet against users like traditional sportsbooks; instead, it collects a fee for matching users with each other. This model ties Kalshi’s revenue to market activity, similar to how a stock exchange profits from trading volume.

Polymarket, by contrast, uses a decentralized, crypto-native model. Today the platform runs on a zero-fee model, signaling that growth and liquidity – not monetization – remain its main priorities. At scale and maturity, one could anticipate Polymarket charging a small take rate.

Across the space, fee structures are evolving to balance liquidity and accessibility. If trading costs are too high, users churn; if they’re too low, liquidity providers have no incentive to participate. The most successful platforms will likely settle somewhere between volume-based exchanges and outcome-based performance models.

One key observation is that sportsbooks can offer significantly larger user incentives because their business model captures structurally higher margins, particularly through parlay bets (a single bet that combines two or more individual wagers into one), compared to the thinner take rates of prediction markets. While sportsbooks do assume outcome risk and carry other operational trade-offs, the disparity in promotional economics is stark from a consumer’s perspective: the absence of large bonuses or incentives on prediction markets is both noticeable and consequential for user acquisition. Prediction markets counter this by having much wider markets and crypto native players (like Polymarket) will likely launch a token as an alternative incentive.

Expanding beyond sports, but struggling to scale
Prediction markets can open doors to categories beyond sports: elections, economics, weather, pop culture, etc. That wide canvas is appealing. But non-sports markets often lack the trading volume, data infrastructure, or natural engagement hooks that sports provide. Monetizing these meaningfully can be a challenge.

Built on crypto vs. fiat
Another key axis of differentiation is whether platforms are built on crypto or fiat rails. Crypto-native markets like Polymarket run fully on-chain, using smart contracts and stablecoins to power trading and settlement. Polymarket has confirmed it plans to launch its own native token, $POLY, designed to incentivize customer acquisition and liquidity across its ecosystem. The token will offer Polymarket a competitive advantage compared to fiat-first players like Kalshi because Polymarket rewards both active traders and market creators while helping bootstrap liquidity in new markets. This on-chain architecture gives Polymarket permissionless reach and global accessibility, but it also brings exposure to volatility, compliance risk, and the constant push-and-pull between decentralization and regulatory clarity.

In addition to Polymarket, some of the names experimenting in the crypto prediction markets space are TryLimitless, Melee Markets, Myriad Markets, XO Market, Clearing Company, and the prediction arm of Crypto.com.

On the fiat side, operators such as Kalshi have chosen to build within the traditional financial stack. Fiat platforms are slower to move but far easier to integrate with mainstream finance.

Ultimately, the split between crypto and fiat prediction markets mirrors the broader dichotomies faced in fintech: openness versus compliance, speed versus stability, experimentation versus institutional trust.

Parlays & house edge vs. flat fees
Traditional sportsbooks lean heavily into parlays and odds architecture, which let them build in a margin (a “house edge”) across multiple outcomes. That’s a high-margin product design built around risk stacking.

In the early days, prediction markets didn’t typically support parlays, which made the experience feel more limited by contrast. But that is changing: Novig already offers parlays, and as of October, Kalshi did as well. That blurs the line even further between “sports bet” and “event contract.”

Where are we headed?

The next phase of prediction markets will stretch far beyond gambling. The same mechanics that make them compelling for sports could make them useful for everything from macroeconomic forecasting to information aggregation.

Sports prediction is the wedge, not the endgame. Once people are trading on the outcome of games, it’s not a stretch to get them trading on the outcome of elections or rate cuts. And when built well, these markets don’t just reflect sentiment – they create it. They function as living, real-time indicators of how people expect the world to evolve.

That’s why economists, regulators, and tech founders are watching them so closely. Prediction markets can serve as information markets: tools for collective intelligence and price discovery. A market on whether inflation will dip below 3% by the end of the year, for instance, becomes a data signal that can inform investors, policy analysts, and even the Federal Reserve. In the same way that futures markets reveal expectations about oil or wheat prices, prediction markets reveal expectations about potentially anything.

There’s also a practical side. These markets could become new hedging tools for individuals and businesses. A company exposed to severe weather, for instance, might hedge risk through a climate-linked market. A content studio could hedge box-office performance or streaming trends. These applications blur the line between speculation and protection, turning betting markets into tools for managing uncertainty. Intercontinental Exchange (ICE) recently announced plans to distribute Polymarket’s data to its customers, enhancing access to sentiment insights on related topics as part of its strategic initiatives and investment in the company.

The infrastructure is also improving quickly. As more platforms gain licenses, APIs open up, and liquidity deepens, prediction markets could become part of the financial web: integrated into brokerages, trading apps, and even social networks. We’re already seeing that trajectory through Robinhood’s integration with Kalshi, Polymarket’s institutional expansion, and PrizePicks’ move into federally licensed derivatives.

In short, the space is transitioning from a curiosity to a foundational layer of the modern attention economy: one where markets and media continuously feed each other.

What This Means for PayPal Ventures

PayPal Ventures hasn’t made a bet in the prediction markets space yet, but we are watching it closely and talking to the companies shaping the space, the start-ups trying to break in, and the underlying rails (payments, clearing, compliance) powering the infrastructure. The energy, capital, and talent flowing in suggest it’s only getting started. In our view, the question isn’t whether prediction markets will scale – they already have. It’s who will own the rails when they do.

In Conclusion

We’re in the early innings of financialized entertainment. Consumers are already conditioned to speculate on stocks, crypto, fantasy lineups, and meme coins. Prediction markets take that instinct and turn it into a product – they make belief tradable. The boundary between finance, media, and gambling is disappearing, replaced by a single, interactive layer of participation.

Prediction markets and gambling are converging into a single behavioral category. Whether users are betting on NFL games, GDP data, or Grammy winners, the underlying impulse is the same: to turn curiosity and conviction into a monetizable position.

For PayPal Ventures, this convergence is both cultural and infrastructural. It’s where consumer attention meets financial plumbing. The companies that succeed here will look like entertainment on the surface but operate like exchanges underneath.

For policymakers, this shift will test the line between gambling law and financial regulation. For investors, it opens a new frontier where infrastructure, liquidity, and behavior collide. And for consumers, it represents a new kind of agency around participation not just in outcomes, but in the very process of how markets interpret the world.