- Multicoin Capital said artificial intelligence is making DeFi software cheaper to build but does not provide the users, capital or proprietary data needed to sustain a business.
- The investment firm identified distribution, capital and proprietary data as DeFi projects’ three main competitive advantages.
- Multicoin Capital expects projects controlling at least one of those resources to capture the most value as financial activity moves onto blockchains.
Multicoin Capital published a report on Nov. 10, 2026, examining how advances in artificial intelligence are changing competition and value distribution in decentralized finance. The firm said AI is sharply reducing product-development costs, but access to customers, capital and specialized data remains difficult to replicate.
Five years ago, launching a decentralized exchange or lending protocol could require about $3 million, five engineers and six months of work, according to the report. Multicoin Capital said tools such as Claude Code can now help build much of such a product over a weekend, while the most valuable parts of the business can take years to establish.
Three competitive advantages
Distribution
Multicoin Capital defined distribution as control over customer relationships, trading flows and user balances. A company with access to a large audience can direct users to different venues and collect a share of the resulting revenue.
The firm cited Hyperliquid, which holds more than $6 billion in USDC used as margin collateral for derivatives trading. Multicoin Capital said control over those balances allowed Hyperliquid to negotiate a share of revenue generated from the stablecoin reserves.
Under an agreement with Coinbase, about 90% of the relevant revenue goes to the protocol’s Assistance Fund. Multicoin Capital estimated that $5.5 billion in USDC could generate about $185 million in annual revenue for Hyperliquid.
Crypto wallets provide another example. MetaMask charges a 0.875% swap fee, while Uniswap’s average fee over the previous 12 months was 0.13%, according to the report. A significant share of Uniswap’s revenue also goes to liquidity providers rather than the protocol itself.
Multicoin Capital said a wallet can therefore earn more than the exchange executing a trade because the wallet controls user access and can route the order elsewhere. However, an interface alone may not provide a durable advantage because users can quickly import a seed phrase into another wallet.
Capital
The second advantage is access to capital needed by traders, market makers and lending protocols. AI can assist with software development, but it cannot independently supply liquidity or absorb losses after financial incidents, the firm said.
Multicoin Capital identified DeFi prime brokerage as one potential use of capital. Such platforms could combine user collateral across venues, provide leverage and finance trading strategies. Traders currently often collateralize positions separately on Hyperliquid, Aave, Lighter, Morpho, Variational and Kamino.
The firm also highlighted risk insurance for DeFi protocols, citing Hyperliquid’s HLP pools and Aave’s Umbrella as mechanisms that help cover losses and bad debt.
Multicoin Capital described a protocol’s own balance sheet, accumulated through retained earnings, as the most durable capital model. Unlike third-party deposits, that capital cannot quickly leave for higher yields elsewhere.
Proprietary data
The third advantage is access to information unavailable on public blockchains and the ability to use it for risk assessment. Multicoin Capital said this is particularly important in lending against and insuring real-world assets.
By contrast, most information needed to assess loans collateralized by bitcoin or Ethereum is public, making the related risk models relatively easy to reproduce with AI, according to the report.
Pressure on blockchain network advantages
Multicoin Capital said layer-1 and layer-2 networks have historically given new projects access to users and capital within their ecosystems. Launching on Ethereum or Solana, for example, allowed a protocol to work immediately with users and assets already on those networks.
Cross-chain interoperability, simpler transfers and fintech companies with millions of customers are gradually weakening that advantage, the report said. AI is also making it easier to move applications between virtual machines and reducing the importance of established developer infrastructure.
As a result, general-purpose blockchains will increasingly compete on transaction execution and settlement speed. Multicoin Capital did not say this would eliminate the value of layer-1 networks, noting that their economic role could change as transaction volumes and fees rise.
Openings for startups
Multicoin Capital said startups can still enter markets that traditional companies avoid because of regulatory, operational or reputational risks. It cited Ethena, which created a synthetic dollar based on delta-neutral hedging and scaled it to $5.4 billion in assets under management, as well as pump.fun and fomo, which developed memecoin trading products and attracted sizable audiences.
The report also pointed to Hyperliquid’s entry into a derivatives market dominated by Binance, Bybit and OKX, including its offer of perpetual futures on companies before their initial public offerings. Project 0 finances DeFi collateral that traditional prime brokers are unwilling to lend against, while USD.ai is developing GPU-backed lending in a segment where banks have limited experience.
Multicoin Capital said startups do not need a durable competitive advantage at launch. They can first target a market that large companies do not serve, acquire customers and build a defensive moat over time.
The firm expects trillions of dollars in financial activity to move onto blockchains over the next decade. It said AI will reduce the cost of building products for that activity, but projects controlling distribution, capital or proprietary data are likely to capture the most value and convert it into stable income for investors.
Source: Incrypted
