Why prediction markets are the largest untapped collateral pool in DeFi

Author: Aly Madhavji

Compiled by: ChainCatcher

Prediction Markets | DeFi Lending | Capital Efficiency | Polymarket | Multi-Asset Collateral | Cross-Margining | On-Chain Credit

TL;DR:

• By the end of 2025, the valuation of prediction markets reached $9 billion (Polymarket) and $11 billion (Kalshi), but the proportion of this capital that can be used for collateralized lending is 0%, resulting in the most extreme capital efficiency gap in the DeFi space.

• The capital utilization rate for token lending can reach 40-80%, NFT lending is about 1%, while prediction market lending is 0%, leading to billions of dollars of digital wealth being effectively trapped.

• The U.S. Commodity Futures Trading Commission (CFTC) approved prediction markets as legitimate derivatives, coupled with a $2 billion investment from the Intercontinental Exchange (ICE), marking institutional recognition that transforms "gambling" into a financeable asset.

• Nettyworth has built an AI-driven infrastructure that unifies tokens, NFTs, and prediction market positions into a single lending interface, aiming to seize first-mover advantage in this undeveloped space.

• The technological barriers remain significant: there is currently no oracle infrastructure for valuing prediction markets, the clearing mechanisms for binary outcome positions remain unresolved, and despite federal-level approval, there is still significant regulatory clarity variance across states.

In October 2025, the Intercontinental Exchange (ICE) placed a $2 billion bet on a company that allows people to wager on everything from presidential elections to Federal Reserve decisions. This funding brought the valuation of Polymarket, which was only five years old, to $8 billion. Two months later, its competitor Kalshi raised $1 billion at an $11 billion valuation. Overall, from January to October 2025, these prediction market platforms processed a cumulative trading volume of up to $28 billion.

However, a seasoned investor holding a $500,000 position in prediction markets faces an absurd reality. This $500,000 is completely idle. No mainstream protocol accepts it as collateral. This capital cannot be used for collateralized lending, cannot be leveraged, and cannot be deployed elsewhere until the position is closed. It can only sit there, locked until the market settles, which could take weeks or even years.

This is not just a minor inconvenience. It represents the most extreme collapse of capital efficiency in the DeFi space, a problem that has remained unresolved for so long because the technological and regulatory infrastructure needed to address it has only just begun to emerge.

When Belief Becomes a Cage

The mechanics of prediction markets are simple. Users buy shares representing a certain outcome. If Donald Trump wins the election, the shares will settle at $1. If he loses, they will settle at $0. Between purchase and settlement, the trading price of these shares reflects the market's probability expectations. A share trading at $0.65 means the market believes there is a 65% chance of that outcome occurring.

Problems arise as large positions accumulate. A fund manager who invests $2 million in the Federal Reserve's interest rate decision six months from now has effectively withdrawn that capital from circulation. Traditional finance solved this issue decades ago through rehypothecation and portfolio margining. Hedge funds can collateralize their entire books for lending, which are viewed as a single risk exposure rather than isolated positions.

But DeFi has not achieved this. Aave's total value locked (TVL) of up to $47 billion is dedicated solely to tokens. GONDI's TVL of over $100 million deals with NFTs. No protocol touches prediction markets, despite Polymarket's open interest peaking at $510 million during the 2024 U.S. election cycle, and according to DefiLlama, there is still $375 million currently.

The gap in capital utilization is staggering. Mainstream DeFi lending protocols have utilization rates of 40-80% on liquid tokens. Despite multiple protocols providing infrastructure, NFT lending barely achieves about 1% utilization in a NFT market exceeding $7 billion. The utilization rate for prediction markets is 0%.

This creates a massive opportunity cost on a large scale. When users must choose between holding positions and obtaining liquidity, they often exit too early. The market becomes inefficient, and the price discovery mechanism is impaired. Funds that should continue to be deployed to reflect true beliefs are frequently traded based on short-term cash needs rather than fundamental views.

Infrastructure Born from Gambling

The regulatory transformation has occurred faster than most observers expected. In 2022, the CFTC fined Polymarket $1.4 million for operating as an unregistered derivatives exchange. By September 2025, the same agency issued a no-action letter, followed by formal approval in November. Acting Chair Caroline Pham publicly stated that the CFTC "must break the past hostility towards innovation" and described prediction markets as "an important new frontier."

Kalshi defeated the CFTC in a legal battle over election contracts, establishing a federal precedent that event contracts constitute legitimate derivatives rather than illegal gambling. This distinction is significant for lending infrastructure. Collateral composed of regulated derivatives can be margin traded, rehypothecated, and financed through traditional banking relationships, which gambling positions cannot achieve.

This timing coincides perfectly with technological empowerment. In December 2025, Kalshi launched tokenized prediction market positions on Solana, creating the on-chain composability needed for DeFi protocols to integrate these assets. Previously, prediction market positions existed in centralized databases inaccessible to smart contracts. Tokenization made them programmable collateral.

Institutional capital follows the clarification of regulations. In addition to ICE's investment in Polymarket, the space has attracted attention from traditional sports betting giants concerned about being disrupted. Market predictions reflect this momentum, with analysts projecting that by 2035, the decentralized prediction market space will reach $95.5 billion, with a compound annual growth rate of 46.8%.

Why Existing Protocols Cannot Fill This Gap

The technical barriers preventing Aave or Compound from simply adding support for prediction markets are far more fundamental than governance votes or protocol upgrades.

Valuation requires different underlying primitives. Homogeneous tokens use real-time oracles from Chainlink or Pyth for price feeds. NFTs need to track floor prices across markets and adjust based on rarity and sales velocity. Prediction markets require probability-weighted valuations that adjust as the likelihood of events changes and account for time decay as the settlement date approaches. Currently, there is no unified oracle infrastructure for such valuations.

The clearing mechanisms are entirely different. Token clearing can occur gradually through Dutch auctions or automated market makers (AMMs). NFT clearing requires matching buyers or accepting significant slippage. Prediction market positions present a unique challenge: they settle as binary outcomes (either $1 or $0), which creates discontinuous value jumps that standard clearing curves cannot accommodate. A position worth $0.65 today could become $1 or $0 tomorrow, with no intermediate state.

Risk correlation becomes exponentially complex. Cross-margining a portfolio that includes ETH, Bored Apes, and Federal Reserve interest rate positions requires calculating how these assets will behave together under market stress. Are they positively correlated (all related to cryptocurrency)? Or negatively correlated (does the prediction position hedge the cryptocurrency risk exposure)? AI models capable of accurately pricing these risks have only recently existed in deployable forms.

The trading volume of NFT lending plummeted 97% from a monthly peak of $1 billion in January 2024 to just $50 million in May 2025, illustrating how quickly the lending market for alternative collateral shrinks when infrastructure proves inadequate. Despite once holding 96% market share, Blend's TVL crashed from $115 million to about $3 million. GONDI emerged as the new leader with a 54% market share, but even that represents only a small fraction of the actively utilized NFT market cap.

Building a Multi-Asset Credit Layer

Nettyworth's approach focuses on underwriting at the portfolio level rather than asset-level approval. The protocol's AI engine analyzes entire wallets while calculating the overall risk encompassing tokens, NFTs, and prediction market positions. This mimics how traditional prime brokers assess hedge fund collateral: as a unified book rather than isolated positions.

The technical implementation relies on audited smart contract infrastructure that holds collateral while loans are active. The non-custodial design means Nettyworth cannot extract user assets; only repayments or effective liquidations can release funds. The AI component operates off-chain, ingesting token price feeds from Chainlink and Pyth, market data for NFTs, and probability assessments for prediction market positions.

Specifically for prediction markets, the protocol simultaneously evaluates current market probabilities and settlement timelines. Positions with six months until the election will be treated differently than those settling next week. Longer timelines imply greater volatility, thus resulting in lower loan-to-value (LTV) ratios. The system also considers the liquidity depth of the underlying prediction market to model the feasibility of liquidation in case of under-collateralization.

Its economic model charges direct borrowers a 2% loan origination fee and a 1% fee for loans initiated through B2B integrations. The revenue-sharing mechanism with integration partners incentivizes wallets, market platforms, and applications to embed Nettyworth's lending engine. This B2B infrastructure strategy emulates Stripe's model: becoming the invisible middleware driving credit functions across the entire ecosystem rather than an isolated platform.

Platform metrics show early adoption signals. As of January 2026, the protocol reported that the wallets it connects hold over $200 million in value, with the system capable of analyzing portfolios containing multiple asset types simultaneously. The company has raised initial funding from the Blockchain Founders Fund, London Real Ventures, Republic, and several other investors focused on DeFi infrastructure.

The leadership team's background spans fintech operations and blockchain security. CEO July Grullon brings deep fintech leadership and a track record of scaling high-growth companies. CTO Ivan Ferrera adds enterprise-level technical expertise, having held IT security positions at BASF and extensive experience in architecting Web3 infrastructure. Advisor Juan Maldonado is a co-founder of Arcus, a payment-as-a-service company acquired by Mastercard in 2021.

Risks Multiply with Ambition

Clarification of federal-level regulations does not guarantee that states will treat them equally. Six states have already issued cease-and-desist orders against Kalshi. In November 2025, a federal judge in Nevada ruled that sports prediction contracts "are not swaps," which could subject them to state gambling regulations, despite federal claims of priority jurisdiction. Discrepancies in circuit court rulings may lead to Supreme Court review, resulting in years of uncertainty.

When combining multiple asset classes, the risks associated with smart contracts multiply. In 2022 alone, up to $403.2 million was stolen through oracle manipulation attacks. Despite undergoing six security audits, Euler Finance still lost $197 million. New multi-asset protocols inherit the foundational risks of DeFi while layering on additional complexities from integrating prediction market oracles, NFT valuation models, and cross-collateral liquidation logic.

Liquidity crises illustrate how quickly collateral markets can freeze. During October 2025, over $19 billion in crypto leverage was liquidated, affecting more than 1.6 million trader accounts. Prediction market positions are event-specific and time-sensitive, lacking secondary market venues for large-scale liquidations.

Adoption friction remains significant. 89% of DeFi newcomers leave protocols within five minutes, with 78% citing unpredictable liquidations as their primary concern. Prediction market lending requires users to understand both prediction market mechanics and DeFi collateral management, creating a compounded complexity barrier that has excluded most retail participants.

Competition from established protocols poses a structural threat. Aave could easily add prediction market collateral through governance votes, leveraging its $47 billion TVL, institutional relationships, and proven infrastructure. If large protocols quickly follow suit, the first-mover advantage in an untested category may be less important than execution advantage.

The Path from Theory to Infrastructure

The convergence of regulatory approval, tokenization on the technical level, and institutional backing has created conditions that did not exist 24 months ago. Prediction markets have transformed from speculative curiosities into derivatives valued in the billions by traditional exchanges. The CFTC has clearly recognized their legitimacy. Mainstream platforms have tokenized positions, enabling their composability with DeFi protocols.

However, the infrastructure lags far behind the opportunity. The idleness of capital is not due to holders lacking belief, but because the financial pipelines needed to mobilize that capital are still under construction. The gap between what is theoretically feasible and what protocols actually support means billions of dollars of trapped liquidity.

Multi-asset lending needs to address issues that single-asset protocols can safely ignore: unified valuation across incompatible primitives, cross-collateral risk modeling, clearing mechanisms for binary outcome positions, and portfolio margining to identify hedged risk exposures. These are not incremental improvements to existing infrastructure. They are the architectural requirements for DeFi capital efficiency to enter its next phase.

Whether Nettyworth or other protocols seize this opportunity, the trajectory of their development is clear. As prediction markets mature from niche applications into a trillion-dollar derivatives market, the infrastructure supporting them must evolve from isolated platforms into integrated financial rails. Otherwise, the consequence will be capital permanently trapped, beliefs turned into cages, and an entire asset class unable to access the foundational financial primitives that keep markets functioning.

The complexity required to build this infrastructure should not be underestimated. The first protocol to successfully address the multi-asset collateral issue at scale will not only serve prediction markets. It will establish a blueprint for how DeFi handles any complex, illiquid, or non-standard asset class. As tokenization expands from prediction markets to real-world assets (RWAs), intellectual property, carbon credits, and other categories that inevitably require on-chain liquidity, this template will become increasingly valuable.

The question is not whether this infrastructure will be built. Capital efficiency demands its emergence. The real question is: which protocol will first solve the hardest technical challenges, which regulatory jurisdiction will prove most inclusive, and whether early entrants can establish network effects before established giants mobilize.

For now, billions remain locked, waiting for the rails that can unlock them.

This article is a reprint of Aly Madhavji's piece published on February 25, 2026, in TheStreet.com, issue 42 of Pione3rs.