By Dean Karakitsos | Founder & CEO, Assymetrix Published: March 2026

Part of the Assymetrix Intelligence Brief series. Previously: Why Prediction Markets Aren't Gambling, The Super Bowl Just Proved Prediction Markets Are Mainstream, The Prediction Market Landscape in 2026, and Wall Street's New Crystal Ball.

Every major technology wave follows the same pattern.

First, vertically integrated companies build everything themselves. Then the market matures, the stack unbundles, and specialized players emerge at each layer. The companies that win long-term aren't necessarily the ones that build the best product at any single layer — they're the ones that own the layer with the most leverage.

Cloud computing did this. AWS didn't win by building the best applications. It won by owning the infrastructure layer that every application depended on.

Fintech did this. Stripe didn't win by building the best bank. It won by owning the payments layer that every fintech company needed.

Prediction markets are doing this right now. And almost nobody is mapping it correctly.

The Stack Nobody Talks About

When Bloomberg, CNBC, or the Wall Street Journal cover prediction markets, they frame it as a platform competition: Kalshi vs. Polymarket vs. DraftKings. Who has more volume. Who has better odds? Who's winning?

That framing is useful but incomplete. It's like covering the early internet by comparing AOL to CompuServe — focused on the portals while missing the infrastructure being built underneath.

The prediction market ecosystem is quietly unbundling into distinct technology layers. Each layer solves a different problem, is controlled by different players, and creates different economic value. Understanding the stack — not just the platforms — is the difference between seeing the industry as a collection of betting apps and seeing it as an emerging financial intelligence infrastructure.

Here's how it breaks down.

Layer 1: Exchanges — Where Contracts Are Created and Traded

This is the layer everyone talks about. Exchanges are where event contracts are listed, priced, and traded. They're the visible face of the industry.

Who owns it: Kalshi ($43B in 2025 volume, CFTC-designated contract market), Polymarket ($33B in 2025 volume, CFTC-approved via QCEX acquisition), and Crypto.com's Derivatives North America (CDNA), which operates as both a standalone exchange and infrastructure provider for DraftKings Predictions and Fanatics Markets.

What they control: Contract creation, order matching, pricing, and the core trading experience. Each exchange defines its own contract language, resolution criteria, and market rules — which is why Kalshi and Polymarket reached opposite conclusions on whether Cardi B "performed" at the Super Bowl, despite both offering markets on the same question.

The economics: Exchanges earn fees on trades, though the models differ. Kalshi charges variable fees depending on contract type. Polymarket currently charges zero trading fees on most markets — a strategy subsidized by its $2 billion ICE investment. This layer is a scale game: the exchange with the most liquidity attracts the most traders, which creates more liquidity. Classic network effects.

The tension: Exchanges are simultaneously the most visible and the most commoditized layer. Contract types are converging across platforms — sports, politics, economics, crypto — and regulatory approvals (CFTC designation) are proliferating. When DraftKings, Robinhood, FanDuel, and Interactive Brokers all offer event contracts, the exchange layer becomes a commodity. The differentiation moves elsewhere in the stack.

Layer 2: Distribution — Where Traders Meet the Market

Distribution is the layer that determines who actually accesses prediction markets. It's increasingly separate from the exchange layer itself — and it may be where the most economic value accrues.

Who owns it: Robinhood (25M+ users, acquired MIAXdx for CFTC licensing), DraftKings (12M+ sports bettors, standalone Predictions app powered by CME and Crypto.com exchanges), FanDuel (50-state coverage via Flutter Entertainment), and Fanatics (150M+ user ecosystem). Polymarket's US waitlist and Kalshi's direct app also serve as distribution, but the trend is clear: distribution is unbundling from exchange operations.

What they control: User acquisition, onboarding, payment processing, and the user experience. Robinhood pipes Kalshi's markets into an interface where prediction contracts sit alongside stocks and crypto — making event contracts feel like just another asset class. DraftKings wraps CME and Crypto.com exchange contracts in a sports-native experience that converts bettors into traders.

The economics: Distribution players take a cut of flow or earn through spread markup. The key metric isn't volume originated — it's users retained. Robinhood doesn't need to build its own exchange. It just needs its 25 million users to trade event contracts the same way they trade fractional shares of Tesla. If even 10% convert, that's 2.5 million new prediction market participants delivered to the ecosystem without a single new exchange being built.

The tension: Distribution is where the sportsbook-versus-financial-instrument debate gets real. DraftKings and FanDuel bring sports bettors. Robinhood and Interactive Brokers bring investors. These are different users with different expectations around regulation, tax treatment, and product experience. The platform that figures out the crossover — converting sports bettors into economic and geopolitical traders, or converting stock traders into event contract participants — wins the distribution layer.

Layer 3: Clearing and Settlement — Where Trust Lives

Clearing and settlement is the invisible layer that most traders never think about — until something goes wrong. It's where contracts are matched, collateral is held, money moves, and outcomes are financially settled.

Who owns it: In the regulated US market, each CFTC-designated contract market (DCM) typically handles its own clearing. Kalshi operates as its own clearinghouse. Crypto.com's CDNA is registered as both a DCM and derivatives clearing organization. On the crypto side, Polymarket settles via USDC on the Polygon blockchain, with smart contracts handling escrow and payout.

What they control: Counterparty risk, margin requirements, payout mechanics, and the speed of settlement. This layer determines whether you get paid when you're right, how quickly, and through what mechanism.

The economics: Clearing is a trust business. It earns by holding collateral and managing risk. As prediction market volumes scale into the hundreds of billions, the clearing layer becomes systemically important. The question of whether prediction market clearinghouses will consolidate — similar to how DTCC consolidated equity clearing — or remain fragmented across exchanges is one of the most consequential infrastructure questions in the industry.

The tension: Blockchain-based settlement (Polymarket's model) and traditional financial clearing (Kalshi's model) represent fundamentally different architectures with different trust assumptions. Blockchain settlement is transparent, instant, and global but requires crypto literacy. Traditional clearing is familiar to institutional capital but slower and opaque. The winning architecture will likely depend on which user base grows faster: crypto-native traders or traditional finance participants entering through Robinhood and DraftKings.

Layer 4: Resolution — Where Truth Is Determined

Resolution is the most underappreciated layer in the entire stack. It's where the outcome of a contract is determined — and it's where the prediction market industry is most vulnerable.

Who owns it: Currently, each exchange handles its own resolution through internal teams. Kalshi employs a "markets team" of lawyers, former traders, and market experts to adjudicate outcomes. Polymarket uses a combination of internal review and its UMA oracle system for disputed contracts. There is no industry-standard resolution authority.

What they control: The definition of truth. When you trade a contract on whether the S&P 500 will drop below 6,200 or whether the Fed will cut rates in June, the resolution layer determines the data source, the precise measurement criteria, and the final verdict. It's the judicial branch of prediction markets.

The economics: Resolution is currently bundled into the exchange layer and treated as a cost center. But the Cardi B debacle — where $57 million was wagered on a question that Kalshi and Polymarket answered differently — exposed resolution as a critical vulnerability. Kalshi settled at "last traded price" (a financial admission of ambiguity). Polymarket initially resolved "yes" and then faced a dispute process. Same event, same question, opposite answers, millions of dollars in the balance. Resolution is where trust is most fragile.

The tension: As prediction markets expand into geopolitics, macroeconomics, and AI milestones, resolution becomes exponentially harder. "Will the US strike Iran by March 31?" is a different kind of question than "Who will win the Super Bowl?" The former requires judgment calls about what constitutes a "strike," what level of military action qualifies, and how to handle partial or ambiguous outcomes. The Maduro contracts showed this in real time — Polymarket initially refused to settle certain Venezuela-related bets even after the capture was announced, because the contract language didn't perfectly match what had occurred.

The prediction market industry will eventually need either a standardized resolution protocol (the "ISDA of prediction markets") or a dedicated resolution layer that serves multiple exchanges. The exchange that figures this out first gains enormous credibility with institutional capital, which cannot tolerate ambiguous settlement.

Layer 5: Data and Intelligence — Where Signal Is Extracted

This is the layer that barely exists yet. And it's the most important one.

Layers 1 through 4 create prediction market data. Layer 5 makes it useful.

The problem: Prediction market data is fragmented across exchanges, inconsistent in format, and lacks cross-platform context. When Kalshi shows S&P 500 correction odds at 58% and Polymarket shows recession probability at 26%, a sophisticated trader needs to synthesize those signals — along with CME rate futures, platform-specific liquidity data, and resolution timeline differences — to form a coherent view.

Today, that synthesis happens manually. A trader opens four browser tabs, compares prices, adjusts for platform-specific biases, and makes a judgment call. This is the equivalent of checking stock prices by calling individual brokers before Bloomberg Terminal consolidated market data in the 1980s.

Who owns it: Almost nobody. This is the gap in the stack. Bloomberg and Dow Jones have partnered with individual platforms (Kalshi and Polymarket respectively) for data feeds, but these are single-platform integrations, not cross-platform intelligence. Some independent dashboards track prices across exchanges, but they're raw data displays without analytical depth — the equivalent of showing stock tickers without any of the analytics, screening, or synthesis that makes market data actionable.

What it should control: Cross-platform price aggregation, spread analysis, liquidity comparison, resolution risk assessment, arbitrage identification, and — most critically — the synthesis of signals across platforms into coherent intelligence that traders and institutions can act on. Not just "what does Kalshi say" or "what does Polymarket say," but "what are prediction markets, collectively, actually saying about this event — and how confident should you be in that signal?"

The economics: Data and intelligence layers historically capture outsized value in financial infrastructure. Bloomberg generates over $10 billion in annual revenue — more than most of the exchanges whose data it aggregates. Refinitiv sold for $27 billion. In every financial market, the intelligence layer that sits above the trading layer eventually becomes more valuable than any single exchange.

The tension: Building this layer requires deep domain expertise in prediction market mechanics — understanding how different exchanges structure contracts, how resolution criteria affect pricing, how liquidity depth varies by platform and event type, and how to normalize data across fundamentally different architectures (blockchain vs. traditional clearing). This isn't a simple API aggregation play. It's an intelligence problem.

Why the Stack Matters Now

Three forces are converging that make the prediction market stack relevant for the first time.

Fragmentation is accelerating. A year ago, there were essentially two prediction market platforms. Today there are ten, with more entering every quarter. DraftKings, Robinhood, FanDuel, Fanatics, Crypto.com, and Interactive Brokers have all entered or announced prediction market products in the last twelve months. Each new entrant adds another data silo, another resolution methodology, another contract format. The fragmentation problem gets worse, not better, as the industry grows.

Institutional capital is arriving. ICE invested $2 billion in Polymarket. Bloomberg and Dow Jones signed data partnerships. CNBC and CNN integrated prediction market data into their coverage. DraftKings is spending $50–200 million on prediction market expansion. This is institutional-scale capital entering an industry with consumer-grade infrastructure. Institutions don't manually check four platforms. They need intelligence layers.

Regulatory clarity is crystallizing. The CFTC has explicitly backed federal jurisdiction over prediction markets. The Supreme Court upheld CFTC authority in the February 2026 tariff ruling. Mick Mulvaney's "Gambling Not Investing" coalition represents the first organized political opposition — which paradoxically validates the industry's significance. Regulatory clarity attracts the institutional capital that demands infrastructure maturity.

These three forces — fragmentation, institutional entry, and regulatory clarity — create the conditions for the stack to unbundle. The vertically integrated era (where each exchange handles everything from contract creation to settlement to data) is ending. The platform era (where specialized players own each layer) is beginning.

Where Assymetrix Fits

We're building at Layer 5.

Assymetrix is the intelligence and synthesis layer for prediction markets. We sit above the fragmentation — aggregating, normalizing, and analyzing cross-platform data to produce the kind of intelligence that no single exchange can provide.

When $529 million flows into Iran strike contracts on Polymarket while Kalshi's economic contracts simultaneously reprice oil, inflation, and Fed rate expectations — the real signal isn't on either platform. It's in the synthesis. That's what we deliver.

The analogy we keep coming back to is Google before PageRank. In the late 1990s, all the information on the internet existed. You could find it if you knew where to look. But the information was fragmented across millions of sites with no unified intelligence layer to rank, synthesize, and surface the most relevant signals. PageRank didn't create the information. It made the information useful.

Prediction markets are at that same inflection point. The data exists. The signal is real. The CEPR study confirmed it across 300,000+ contracts. But the data is scattered across platforms with different architectures, different contract structures, different resolution criteria, and different trader bases. Making that data useful at scale — for traders, for institutions, for anyone who needs to understand what prediction markets are collectively saying — requires a dedicated intelligence layer.

That's what we're building. And in a stack that's unbundling fast, we believe it's the layer that matters most.

At Assymetrix, we're building intelligence at the intersection of news, events, and technology — synthesizing real-time data across every prediction market platform so event-driven traders see the full picture. Follow us for more industry analysis.