PANews published the essay this article rewrites on November 11, 2026, at 10:29 a.m. China Standard Time (UTC+8). It was written by Spencer Applebaum, an investment partner at Multicoin Capital, and compiled by 大钳子. What follows is that essay’s argument. It is not a finding that any firm named in it is solvent, audited, or legally obliged to make customers whole, and it is not an investment recommendation.
What got cheap, and what did not
Applebaum’s opening contrast is operational. Five years before the essay, building a decentralized exchange or a lending protocol meant raising about $3 million, hiring five engineers, and spending six months on a minimum viable product. A founder using Claude Code, he writes, can now do most of that work in a weekend. Ask a frontier model for a perpetual-futures exchange and it can produce a matching engine, a risk engine, a trading interface, an API, and most of the basic functions.
Financial services, in his view, remain among the hardest businesses to commoditize, because the valuable parts have little to do with the software. A loyal customer base, a large balance sheet, and proprietary data plus the ability to assess risk take years of relationships, a track record, and trust. AI can cut the cost of writing software. It does not make those easier to copy.
A model can write a lending protocol in an afternoon. It cannot place $10 billion of stablecoin deposits into the contract, and it cannot bring 28 million customers to a newly formed digital bank.
From that split, Multicoin is constructive on DeFi and narrow about who gets paid. Over the next decade, the essay expects trillions of dollars of financial activity to move on-chain, and AI to cut the cost of the products that serve it. The resources that decide a financial firm’s value, and its barriers to competition, do not become easier to reproduce.
How the firm got to this frame
Multicoin has revised its view of DeFi value capture as the market changed. Early on, it focused on application tokens capturing value through work tokens and burn-and-mint equilibrium designs. It then shifted toward defensibility and what it called unforkable state: anyone can copy open-source code, but not liquidity and network effects. More recently it argued that a DeFi protocol captures more value only when it takes on and manages some form of risk. Applebaum says each of those views has been borne out to some degree. This essay extends them.
He also states a broader belief: DeFi will keep taking share from centralized finance and will eventually cover traditional finance as a whole. DEX share relative to centralized exchanges, he writes, is already at a historical high and still growing, citing The Block and DefiLlama. The essay does not state a percentage.
Tokenized assets, he adds, are starting to see real adoption, whether U.S. Treasuries, private credit, or equities. The growth is no longer driven only by crypto-native users. Traditional financial institutions, fintech companies, institutional asset managers, and traditional brokerages are building on-chain products and bringing their own customers, with rwa.xyz cited as the source of that picture. The essay text does not attach a market-size figure.
The warning attached to the growth is that today’s leading DeFi protocols are not guaranteed to keep it. As the cost of software and of data analysis falls, value moves toward whoever holds scarce resources that software cannot create.
The essay names three moats: distribution, capital, and proprietary data. A fourth factor, risk tolerance, works differently. Applebaum treats it as a way into a market and returns to it at the end.
Distribution: who can redirect the balance
Distribution, as defined here, is control over where order flow or fund balances go. It is not all equally sticky. An account that is already funded and hard to move can be a real barrier. A user interface alone is easier to contest.
The essay sets this against Joel Monegro’s 2016 fat-protocols thesis. That argument held that most internet value accrued at the application layer, at firms such as Google, Facebook, and Amazon, while blockchains would invert the relationship and let base protocols capture most of the value. Applebaum says events are moving the other way, and faster. As protocols get cheaper to build, value shifts toward whoever holds the customer relationship and can send users and funds to the venue that pays the most. Sometimes that party is an application. Sometimes, as with Hyperliquid, it is a blockchain that is also an application.
The stablecoin stack, and who does not hold the reserve
The stablecoin value chain is the essay’s clearest case.
Historically, once an issuer is large enough and rates are not near zero, issuing a stablecoin has been a very profitable business. Tether and Circle hold dollar reserves and earn interest on them. In most cases, the essay says, they do not pay that interest to holders. Tether alone, since 2021, has earned more than $20 billion of interest income at margins near 100 percent, on the essay’s reading of DefiLlama.
That income is the issuer’s, generated by reserves the issuer holds. It is not evidence about a trading venue’s own assets, its liabilities, or what it owes customers.
Hyperliquid is the bargaining example. More than $6 billion of USDC sits on the platform as collateral for derivatives. Those balances produce substantial reserve yield for Circle and its distribution partners. Hyperliquid’s claim is that its platform caused the balances to exist, and therefore has leverage: if Circle will not share the yield, Hyperliquid can substitute a stablecoin that will, and in theory could issue one itself.
Earlier in 2026, through its Aligned Quote Asset (AQA) framework, Hyperliquid reached an arrangement with Coinbase. Under it, about 90 percent of the USDC reserve yield on the platform flows to the protocol’s Assistance Fund and is used to buy HYPE, rather than remaining with Circle or Coinbase. Applied to $5.5 billion of USDC in the treasury, that 90 percent share is described as about $185 million of revenue for Hyperliquid at the time of the essay.
The custody and license chain is unchanged by that split. Hyperliquid does not run minting and redemption, does not manage the treasury, and does not hold the regulatory license. What it holds is the distribution channel. The essay’s point is that this has been enough to demand a larger share of the economics. A contractual claim on yield is not ownership of the reserve, a match of customer liabilities, or proof that either party can meet withdrawals under stress.
Wallets and frontends keep the relationship, not the engine
Wallets and frontends are the other distribution case.
Early non-custodial wallets, the essay says, struggled to monetize. They later recognized that trading was crypto’s most profitable business and, a few years before this piece, made trading a core product.
On fees, MetaMask charges 0.875 percent. Uniswap’s average trading fee over the prior 12 months was 0.13 percent, and most of Uniswap’s fee goes to liquidity providers rather than to the protocol. The wallet earns far more from the trade than the venue that executes it.
The venue has to attract liquidity providers, maintain complex infrastructure, and compete on execution quality. MetaMask has to win and keep the customer. If a DEX is not competitive on price, MetaMask can route the trade elsewhere. As its order flow grows, it can internalize more of the flow and keep more of the take.
Perpetuals are starting to look similar. Phantom lets users trade perpetuals through Hyperliquid and charges an extra 5 basis points on top of the 4.5 basis points Hyperliquid charges. Hyperliquid runs the risk engine, the matching system, and the liquidity. Phantom, holding the interface and the customer relationship, earns more.
The essay is explicit about the limit. Builder teams in the Hyperliquid ecosystem are already in a price war, cutting fees to hold users. Even a frontend can be commoditized. Crypto wallets are also fragile because leaving is cheap: import a private key or seed phrase into another wallet, and the same funds are available through a new interface within minutes. Leaving a bank or a fintech service is not. The user opens a new account, completes identity checks, moves money over traditional payment rails, relinks payroll, and may have to update other connected services.
A fintech with 28 million funded accounts
Robinhood, in the essay, has 28 million funded accounts. Uniswap is the main DEX on Robinhood Chain, and Robinhood has not objected.
Applebaum’s point is not a prediction that Robinhood will replace it. He notes that, historically, owning liquidity has not been Robinhood’s primary goal. The claim is narrower: as AI lowers the cost of financial software, nothing technical stops Robinhood from building a DEX and directing users to it. With that reach and current programming tools, it could, nearly overnight, launch one of the largest layer-2 DEXs in DeFi. The essay does say the company has already begun shifting substantial trading flow from Kalshi to its own prediction market, Rothera.
What he wants emphasized is the sequence. Building the protocol is no longer the hard part. Attracting users, capital, and proprietary data is.
The allocation chain he draws runs upward. A frontend such as Phantom takes more of the economics from the trading venue. The venue, such as Hyperliquid, takes more from the stablecoin issuer, such as Circle. At each step, the party that controls distribution has more pricing power over the infrastructure underneath. How much of that pricing power survives depends on how easily users leave. Hyperliquid obtained about 90 percent of the reserve yield in the arrangement above. Frontends in its ecosystem are already competing fees down toward single-digit basis points.
Capital: whose money, and who bears the loss
A protocol or a piece of back-end infrastructure without a direct user entrance is, on this account, at a disadvantage. Applebaum does not stop there. Capital is a separate moat, and in his telling one of the most important.
Capital is what the market runs on. Traders need it for leverage. Market makers need it to keep quoting. Lending protocols need deposits before they can lend. AI can write code and automate workflows. It cannot create the funds. When something fails — the essay cites Aave’s KelpDAO episode and MakerDAO’s Black Thursday — someone has to bear the loss. A visible on-chain balance does not identify that someone, and it does not measure the liability against the balance.
Liquidity that can leave
He treats liquidity as a logarithmic network effect: traders go where liquidity is deep, and liquidity follows traders. Hyperliquid’s open interest is a real competitive advantage, but the capital belongs to users, who can move it. DeFi liquidity has already moved from protocol to protocol, often pulled by token incentives and so-called vampire attacks, and a large share has left once the incentives stopped.
Two forms of capital, he argues, can support a longer defense.
Financing positions that offset
Prime brokers are central in traditional markets. They provide leverage, finance trades, and let a hedge fund use one pool of capital across markets and strategies instead of posting collateral at every venue. The business is large. Pushed by growth in hedge funds and by financing-heavy multi-manager pod-shop models, industry prime-brokerage revenue rose from about $15 billion in 2020 to about $27 billion in 2024.
Multicoin thinks the DeFi version could be larger still. Traditional prime brokers mainly serve large institutions, because serving small clients costs too much to operate. Smart contracts can automate much of that work and extend the service from a person trading $500 to a team managing hundreds of millions of dollars at Millennium.
Today, a trader generally posts collateral separately on venues such as Hyperliquid, Aave, Lighter, Morpho, Variational, and Kamino. That uses capital inefficiently, and the essay expects the problem to worsen as tokenized equities, U.S. Treasuries, private credit, and other real-world assets arrive on-chain. A fund that is long tokenized equities on one platform and short an equity-index perpetual on another should not, where the risks offset, be required to post full margin against both.
Writing the software to administer that book is getting easier. Having a balance sheet large enough to finance the positions, access to low-cost funding, and risk management that avoids large losses is not. Over time, a prime broker with cheaper money can offer better terms, attract more clients, and use scale to cheapen its funding further. Applebaum expects meaningful economies of scale.
The scarce piece is not the margin engine. It is who has the money to lend, on what terms, and who absorbs the loss if the hedge does not offset as modeled.
Capital that stays when the incentive ends
The second use of capital is backstopping risk. This connects to an earlier Multicoin argument on DeFi DAO risk management: a protocol that manages risk has an associated token that is harder to displace with a fork than a protocol that does not. Hyperliquid’s HLP vault is the example. It supplies liquidity by making markets and taking on liquidated positions. Depositors share the profit and must also take the loss. In March 2025, HLP lost about $4 million while closing a large ETH position. The later JELLYJELLY episode exposed it to further risk. Both, the essay says, show why an insurance fund at a derivatives DEX matters.
Aave’s Umbrella is described as a similar arrangement: users backstop bad debt and receive a return for doing so.
Anyone can copy the HLP or Umbrella contracts. They cannot copy the capital inside them. The catch is that this capital comes from third parties. If depositors see a higher yield elsewhere, they can leave. One reply, offered as a counterpoint rather than a settled fact, is that some capital is stickier because token holders and depositors overlap. A HYPE holder might leave dollars in HLP to make the venue safer and to protect the token position.
The form Applebaum calls most durable is a balance sheet the protocol builds from retained earnings. That capital does not depart because a newer yield or a points program appears.
Binance is the recent illustration, and the essay limits how far to read it. After an October 29 hack that cost $6.3 billion, presented as a recent case, Binance covered the loss with its SAFU reserve fund and with company funds, mainly the latter, and reopened withdrawals within days. Net assets that left amounted to only 3 percent of total assets. Applebaum writes that it is still too early to conclude, but that early evidence suggests trust in the strength of the balance sheet may have helped avoid a bank run.
That report does not establish solvency, the size of remaining customer liabilities, the quality of an audit, or a legal protection for customers. A reserve used to fill a hole, and a net outflow of 3 percent, are the response as the essay states them. They are not a proof built from wallet balances, and they do not say what was still owed after the top-up.
As DeFi moves further into private credit, tokenized equities, and other real-world assets, the essay expects demand for insurance and risk capital to keep rising. Its conclusion on this moat is that protocols that actually supply and manage capital will capture more value than protocols that only supply software.
Proprietary data: public prices are not an underwriting file
The third moat is information, with a premise the essay flags. DeFi was built on the idea that financial data should be public. That works for crypto-collateralized lending and for spot trading. As more real-world assets move on-chain, more businesses will be built around information that is not public. The moat is proprietary data together with the ability to assess risk using it.
Curators managing $9.5 billion, at thin fees
Lending shows the pattern most clearly. Curators on Morpho, Kamino, and Euler assess risk and manage depositor funds totaling $9.5 billion. They hold limited control inside the vault contracts: which lending markets receive capital, what the risk parameters are, and monitoring of collateral.
Despite that $9.5 billion, the income they generate is very limited, especially next to traditional asset managers with wider discretion and responsibility. Many managers charge no fee. The largest DeFi risk curator by assets, Steakhouse Financial, charges a management fee of only 22 basis points. The average across managers is 19 basis points.
Applebaum gives two partial reasons, then a larger one. Risk-adjusted yields in DeFi are not attractive, which may relate in part to idle capital on Aave. Morpho and Kamino have also commoditized the vault layer and sit between the risk manager and the depositor, which is distribution again. The larger problem is that the risk managers are mostly looking at the same data. If the collateral is ETH or BTC, the price, the liquidation rules, and the collateral facts are public. AI makes a risk model easy to copy, so curators compete on judgment, including a risk appetite much of which is already set by parameters, and on track record. Over a long enough period, he expects that to press fees down.
What a block explorer does not contain
OnRe is the contrast. It is an on-chain reinsurer. It raises capital from investors and uses that capital to underwrite real-world insurance risk. Doing so requires historical claims data, loss probabilities, and the insurance contracts themselves, plus people who understand that material. None of it is on a block explorer. A frontier model with no access to the reinsurance data cannot do the underwriting, however capable the model is.
That, the essay says, is why OnRe can charge a 75-basis-point management fee plus 20 percent of gains. At then-current returns, annualized fees come to about 3 percent, roughly 16 times the average fee of DeFi risk managers.
These are different risk businesses. Part of the fee premium reflects illiquid assets and the risk of loss. Applebaum treats that as the point: a firm with proprietary data, underwriting skill, and customer relationships can charge a clearly higher fee. If the data and the tools are the same as everyone else’s, the business is on the way to becoming a commodity.
Chains that used to supply two of the three moats
If DeFi grows for a long time, who can charge, and who will find charging harder?
One pessimistic reading in the essay is that layer-1 and layer-2 infrastructure tokens are in a difficult position. Blockchains used to hand new DeFi applications two of the three moats, distribution and capital.
Distribution, historically, came from users leaving assets on the chain where they were most active. An application on Ethereum reached Ethereum users; an application on Solana reached Solana users. If Uniswap, Compound, Aave, Kamino, or MakerDAO were created from zero today, that ecosystem advantage would matter, because the builder would have no existing customers, assets, or financial relationships. Starting where users had already gathered made the company easier to begin. Interoperability, Applebaum writes, is wearing that advantage down. Users can enter through an application and trade assets on almost any chain while barely considering the chain underneath. More new users will also arrive through fintech companies such as Revolut, Nubank, and Robinhood, which already have tens of millions of customers and do not require a wallet to be funded first.
Capital used to arrive through a chain’s wealth effect. As SOL or ETH rose, users had more to deploy into applications on that chain. Bridging was difficult, and users often felt a strong attachment to one chain, so capital tended to stay. Both conditions have weakened. Cross-chain movement is easier, and capital now enters as stablecoins, tokenized assets, and institutional money. New protocols depend less on the wealth effect of any single token. Issuers of real-world assets also want the same asset on every chain. Securitize, Superstate, and Ondo are the examples cited.
Data was never much of a chain-level moat. Nearly all on-chain financial information is public, so neither the chain nor the application held much proprietary data it could use as a barrier. AI is also reducing another historical advantage, the developer ecosystem. A large part of Ethereum’s edge over newer chains such as Solana or Aptos was mature tooling, audited code, and dense integrations. As AI makes it easier to move an application between virtual machines, that edge matters less.
General-purpose blockchains are therefore losing the distribution advantage and the capital advantage, and they did not truly build a data moat. The implication is that they may be left competing mainly on execution and settlement, where blockspace is plentiful and fees keep falling.
That is not a claim that a blockchain has no value. Applebaum points to an argument Tushar made in a summit speech several years earlier: a layer-1 token can at the same time be a capital asset, a commodity, and a store of value. If Solana eventually processed one billion payment transactions a day and charged $0.01 each, annual revenue would be $3.65 billion. The essay says volume was still growing, citing Blockworks Research. It does not treat that fee figure as current revenue. The conclusion it does draw is that the source of layer-1 and layer-2 value may differ entirely from what investors first expected.
What an application has to hold
For applications, the firms that capture the most need at least one of the three:
- Distribution: a direct customer relationship that is hard to leave. Robinhood and Hyperliquid are the examples.
- Capital: a balance sheet that can bear losses.
- Information: the ability to underwrite risk with data other participants cannot obtain.
However strong the model becomes, those three are much harder to copy than the software.
Multicoin does not treat equity versus a token as the important question. Both, it says, can capture value. Recent SEC guidance and rule proposals, which the essay does not name, have opened a path for tokens to capture value in a compliant way, and the firm says it intends to invest in opportunities there. What matters is whether the business, in either form, controls a scarce resource, has pricing power because of it, and can turn that into a return for its investors.
As more financial activity moves on-chain, the essay expects the gap to widen between protocols with a moat and protocols that have only software. Describing itself as one of the largest investors in the sector, Multicoin says it spends substantial time on which projects across the stack actually hold these moats and can keep them.
The incumbent advantage, and an opening that is not a moat
It is reasonable, Applebaum grants, to read the argument as favoring firms that already exist. Robinhood has the customers. Goldman Sachs may have more capital than any DeFi prime broker. Moody’s has the data and the related licenses. If writing software is becoming close to free, and the real moats are expensive resources, the holders of those resources have the advantage.
Startups still have one advantage in this account: they will take risks the large firms will not. AI cuts both ways. The same tools that would let Robinhood ship a DEX in a weekend would let a three-person company do it. As the cost of a competitive product falls, the opening is a market a giant will not enter and a customer that giant is not serving.
The essay treats several cases as already visible.
Ethena built a synthetic dollar around a delta-neutral hedge. Traditional institutions know the strategy, a basis trade, but had been unable or unwilling to bring it onto crypto exchanges. Ethena’s assets under management had reached $5.4 billion.
Pump.fun and fomo built new trading experiences around meme coins and used that category to establish large distribution. The essay calls them among the most successful projects of this market cycle.
Hyperliquid entered a derivatives market dominated by Binance, Bybit, and OKX and took share quickly. It also began listing perpetuals on pre-listing companies such as OpenAI and Anthropic, markets traditional exchanges would not enter. Those markets drew new traders and liquidity. The essay points to wallets created on the day of the SpaceX IPO, citing HyperTracker, as a sign of that demand. The text does not state the wallet count.
On capital, Project 0 finances DeFi collateral that traditional prime brokers will not accept. On data, USD.ai underwrites loans against GPUs, collateral banks have almost no experience with.
What they share is entry into markets established firms were not serving. The reason might be regulatory limits, operational complexity, reputational risk, or an opportunity that looked too small. While that remains true, a startup can still enter on asymmetric terms. Crypto’s history has shown this more than once, and the essay expects it to continue.
That is also why Applebaum finds AI and DeFi together attractive. AI is sharply lowering the cost of building financial products. DeFi lets a startup reach customers and capital directly, without first reconstructing an entire financial system.
Risk tolerance is not itself a lasting moat. Regulation catches up, and large firms enter once a market has been shown to work. Spot bitcoin is the example. A startup does not need a moat on the day it is founded. It needs an entry incumbents will not take, and enough time to turn that entry into one or more of the three moats. After that, it can compound the advantage.
Where the essay lands
Over the following decade, Applebaum expects trillions of dollars of financial activity to move on-chain, and AI to make the products serving that activity easier to build than ever. Software can become a commodity. The elements that decide what a financial firm is worth have not changed. Customers, capital, and proprietary data remain hard to obtain, and the firms that hold them have pricing power. That is why Multicoin says it is constructive on the next generation of DeFi businesses.
As AI development tools spread, a new company can compete with a large firm faster and at lower cost than before. If it then establishes one of the moats, it has a chance at an unusually durable advantage. Code can be copied, or regenerated with the latest model. On this account, the moats cannot be generated that way. That is the opportunity Multicoin says it sees as a crypto investor.