Multicoin Capital 合夥人 Spencer Applebaum：DeFi 2.0，資產終於追上基礎設施

0:00DeFi 1.0 的誕生與設計邏輯

0:00大家好，我是 Spencer Applebaum，Multicoin Capital 的普通合夥人暨創投共同負責人。幾週前，我和同事 Shayon 發表了一篇名為「DeFi 2.0」的文章，今天我會帶大家一起深入探討。其實我們的核心論點很簡單：第一代 DeFi 所建構的基礎設施，都是圍繞著加密資產打造的，像是 Bitcoin、ETH、SOL、HYPE 等等。但現在，國庫券、股票、信貸市場、外匯等資產也開始上鏈。隨著資產本身的變化，底層的金融基礎設施也必須跟著調整。Hi everyone, my name is Spencer Applebaum, and I'm a general partner and the co-head of venture at Multicoin Capital. A few weeks ago, my colleague Shayon and I published an essay that we call DeFi 2.0, and that's what I'm going to walk through today. The basic thesis is actually pretty simple. All of the infrastructure that we built in the first generation of DeFi was built around crypto assets, so things like Bitcoin, ETH, SOL, HYPE, etc. But now we have treasuries, equities, credit markets, FX. All of these things are starting to move on-chain. And so as the assets themselves change, the financial infrastructure underneath them needs to change as well.

0:46今天這場分享會分成三個部分：第一，為什麼 DeFi 1.0 會長成現在這個樣子；第二，當現實世界資產（RWA）上鏈後會帶來哪些改變；第三，這些變化會為我們和其他人創造哪些投資機會。先聲明一下，法律團隊要我提醒大家，今天的內容都不是投資建議，也不是 Multicoin 的任何服務邀約。So this talk is going to be in three parts today. First, why DeFi 1.0 looks the way that it does. Second, what will change as real-world assets come on chain. And then third, where we think that creates investment opportunities for us and others. Before we dive in, my legal team asked me to tell you that nothing you'll hear or see today is financial advice or an offer or solicitation for Multicoin's advisory services.

1:15DeFi 1.0 為何與傳統金融不同

1:15如果你來自傳統金融，回頭看 DeFi 早期出現的核心產品，會覺得這些設計很奇怪、效率很低。舉例來說，AMM（自動造市商）讓資金被動地放在曲線上，很多時候成交品質也不佳。過度抵押的借貸協議，要求你拿 150 美元抵押，才能借出 100 美元。借貸利率可能幾天內就從 5% 變到 20%。另外，傳統金融有到期日的期貨，但加密產業卻流行永續合約這種永不結算的產品。乍看之下，好像 DeFi 只是把金融重新做了一遍，還做得更差。但我認為這個結論是錯的。If you come from TradFi and look at the core products that emerged in DeFi pretty early on, a lot of them look pretty strange and candidly pretty inefficient. AMMs, for example, passive liquidity that sits on a curve, and in a lot of cases they don't really offer great execution. Over-collateralized lending asks you to put up $150 to borrow $100. On borrow-lend protocols, rates can swing from 5 to 20% in just a matter of days. And then lastly, instead of dated futures, the crypto industry converged around this perps product that never expires. From the outside, that conclusion might be that DeFi just rebuilt finance, but in a worse way. But I actually think that's the wrong takeaway.

2:00這些產品之所以看起來怪，是因為它們根本不是為國庫券、股票、傳統資產或資產管理人設計的，而是為加密資產、Bitcoin、ETH、SOL，甚至是數千種長尾代幣量身打造。這些資產本質上就完全不同。例如，這些代幣的年化波動率高達 50% 到 100%，而傳統資產只有 5% 到 15%。加密資產通常沒有現金流，也沒有到期日。借貸對手方往往是匿名錢包，而不是有信用紀錄的機構。交易執行面也多是散戶小額交易，而非機構大額交易。These products look strange because they weren't actually built for traditional assets like a treasury or an equity or, you know, even traditional market participants like asset managers. They were built, again, for crypto assets, Bitcoin, ETH, SOL, and even more so thousands of long-tail tokens, and those are just fundamentally different assets. For example, these assets that we know as tokens have 50 to 100% annualized volatility instead of the, call it, 5 to 15% on the traditional side. Crypto assets generally don't have cash flows, and they don't mature. It's often wallets that are pseudonymous that are borrowing instead of having known counterparties where you can actually call margin after the fact. And then on the execution side, a lot of it is, you know, generally retail trading small lots rather than institutions trading institutional-size blocks.

2:54所以從這個前提出發，DeFi 1.0 的設計就變得合理了。我們來看幾個例子。以 AMM 為例，如果你在交易 Apple 股票或歐元兌美元，這些市場早就有一堆造市商在競爭，價差只剩下幾個基點。但如果你今天下午在 pump.fun 發行一個新代幣，或在 MetaDAO 募資，問題不是怎麼把價差從 20 個基點壓到 10 個基點，而是怎麼讓市場本身成立。AMM 幫這個產業解決了「市場不存在」的問題。你只要把兩種資產放進池子，馬上就有價格發現和流動性。沒錯，AMM 資本效率很差，但資本效率根本不是這些產品要解決的問題。So once you start from that premise, almost everything about DeFi 1.0 actually does start to make sense, and we're going to talk about a couple of examples. Let's start with AMMs. If I'm trading Apple stock, or, you know, Euro USD. The market for that stuff already exists, right? There's dozens of market makers out there competing over what is effectively basis points. But if I launch a token on pump.fun this afternoon, or I raise a small round on MetaDAO, for example, my problem isn't really getting the spread down from 20 basis points to 10. The problem is getting a market to exist at all. And AMMs have solved that for this industry pretty well. You have two assets in a pool, you immediately have price discovery and you can start accruing some liquidity. So yes, obviously AMMs are capital inefficient, but capital efficiency was never really the problem that these things needed to solve.

3:47借貸協議也是同樣的道理。匿名錢包沒有財報、沒有信用紀錄，你沒辦法審核借款人，只能審核抵押品。像 Galaxy 的這張圖，鏈上 USDC 借貸利率歷史上曾經從 2% 飆到 15% 又回來，這在傳統金融是不可接受的。但如果你是用 2,000 萬美元的 SOL 抵押借 1,000 萬 USDC，SOL 明天可能就漲跌 10%，誰還在乎年利率多變幾個百分點？永續合約也是同樣邏輯，因為加密資產沒有到期日，交易者只想要持續槓桿曝險，所以永續合約才會這麼受歡迎。總結來說，DeFi 1.0 並不是傳統金融的「半成品」，而是針對加密資產特性最佳化的設計。Lending's also the same story, right? An anonymous wallet has no income statement, it has no credit history, you can't underwrite the borrower, and so you have to underwrite the collateral. If you look at this chart from Galaxy, on-chain USDC borrow rates have gone from, you know, 2 to over 15% and back several times throughout history. In TradFi, we think that's probably unacceptable. But again, if you're borrowing, let's say, $10 million of USDC against $20 million of SOL, you probably don't care because the SOL might move 10% tomorrow. So who cares about the annual interest rate changing a couple percent? Perps actually follow the same logic. So crypto assets don't mature, and traders generally just want continuous leverage exposure to whatever the underlying is, and that's why perps have become so popular. I think the important TLDR from the last two slides is DeFi 1.0, it wasn't really an immature version of TradFi. It was actually pretty well optimized for the crypto assets that people were trading.

4:47RWA 上鏈：資產結構與基礎設施的變革

4:47接下來這張投影片才是重點。大家應該都同意，DeFi 底層的資產結構正在改變。國庫券、股票已經上鏈，信貸、商品、外匯等也很快會跟進。這個機會的規模非常巨大。傳統金融市場每天交易約 20 兆美元，而加密市場只有三分之一兆。到目前為止，RWA 的討論大多集中在「如何把資產上鏈」這件事，但我們認為更有趣的問題是：這些資產真的上鏈之後，會發生什麼事？因為這些資產和傳統加密資產的行為差異很大，我們預期會出現一整套全新的金融原生模組（primitives）。So this slide is the one where I think things actually get interesting. So I think we can all agree, that's probably why we're here, that the asset base underneath DeFi is changing. Treasuries and equities are already moving on-chain, and credit, commodities, FX, etc., those are probably next. And the scale of the opportunity, importantly, here is really enormous. TradFi markets trade roughly 20 trillion dollars a day, whereas crypto trades about a third of a trillion. In our view, thus far most of the RWA conversation has been about tokenization itself, right? Meaning, how do we actually get these assets onto a blockchain? But we think the more interesting question is what happens once those assets actually do get there, because these assets behave so differently from what we've known as traditional crypto assets, we expect a whole new set of primitives to emerge.

5:40我們認為 RWA 和加密資產有五大差異：第一，RWA 波動率低、流動性高、價差小；第二，很多 RWA 有存續期間和現金流；第三，RWA 通常是更好的抵押品；第四，市場參與者多為機構，且可以識別身分；第五，未來鏈上投資組合會跨足多種資產類別。接下來我會逐一說明這些差異，並說明每當有新資產類型出現，原本在 DeFi 1.0 看似多餘的金融模組，忽然就變得很有必要。So we think that RWAs differ from crypto assets in five broad categories. So first, RWAs generally have lower volatility, and they have tighter spreads because they're more liquid. Second, a lot of them have duration and cash flows. Third, they usually, generally speaking, make for better collateral. And the market participants, i.e. the borrowers, you can actually identify. Fourth, much larger institutional participants are already trading these assets. And then fifth, we're going to have portfolios on-chain that span lots of different asset classes. So in the next few slides, I'm going to take these one at a time. And I think the pattern to watch for is every time you have a new asset, a primitive that previously looked unnecessary in DeFi 1.0 suddenly starts to make a lot of sense.

6:32波動率、流動性與市場結構的轉變

6:32第一個差異是波動率和價差。你買 meme coin 吃 2% 滑價，可能根本不在乎，因為一小時後幣價可能就漲跌 30%。但如果你在交易國庫券或外匯，10 個基點的價差就足以決定整筆交易的利潤。因此，市場微結構開始變得非常重要。這也改變了你需要的流動性提供方式——你不再需要被動資金放在 AMM 曲線上，而是需要專業造市商根據其他市場即時調整價格。所以我們認為訂單簿、RFQ 系統，以及像 Solana 上出現的 Prop AMM 會越來越重要。這些市場對造市商來說也更容易啟動，因為他們本來就在傳統金融交易這些資產，也有很多傳統避險管道。So the first difference is volatility and with it how tight spreads are. If you buy a meme coin and you eat 2% slippage, you probably don't really care, right? The token might be up or down 30% an hour later. But if you're trading treasuries or FX, for example, 10 bips can literally be the economics of the entire trade. And so market microstructure starts to meaningfully matter. And so that changes the type of liquidity provisioning you actually want, right? You don't want passive liquidity sitting on an AMM curve. You want professional market makers constantly repricing off of where the asset is trading on other venues. And so that's why we think order books, RFQ systems, and prop AMMs that we're starting to see on Solana are becoming so much more important. These markets, importantly, are actually a lot easier to bootstrap from a market maker perspective because they're already trading these assets in TradFi, and there's plenty of traditional venues that they can hedge on.

7:28很多人說 DeFi 還沒準備好迎接這些新型態的市場結構，但你看這張投影片的資料就知道，像 Hyperliquid、Lighter 在百萬美元訂單上的滑價，已經和 Binance 差不多，甚至有時 Hyperliquid 表現更好。Solana 上的 Prop AMM 也能提供比 Binance 更緊的 SOL/USDC 流動性。換句話說，DeFi 其實已經準備好迎接更進階的市場微結構。I know that there's a lot of commentary out there that DeFi is not ready for these primitives, but if you just look at this data on the slide, right, on a million-dollar order, Hyperliquid and Lighter are within a few basis points of Binance, and actually Hyperliquid is quite competitive. In some cases, they actually do offer better execution. And then on Solana, Prop AMMs now offer tighter SOL USDC liquidity than Binance, which is the most liquid market on Solana. So said another way, DeFi is actually ready for prime time in these more advanced market microstructure primitives.

8:02選擇權、存續期間與現金流產品的崛起

8:02低波動率也會改變選擇權市場。幾年前，像 Ribbon、Katana、Friktion 等團隊在 Ethereum、Solana 等鏈上做了 covered call vault（備兌買權金庫），讓你可以存入 SOL、ETH、Bitcoin，賣出買權收取權利金。這種產品在傳統金融行之有年，但在加密資產上卻不吸引人，因為 SOL、ETH 這些幣一個月內可能漲跌 50%，你為了幾個百分點的權利金就放棄所有上漲空間，根本不划算。但如果把同樣的產品用在一年只漲 10% 的股票，covered call 就很有意義了，其他像 collar、下檔保護等結構型商品也會興起。這些選擇權金庫不是錯，只是用錯資產、用錯時機。Lower volatility also changes options, right? Several years ago. There were many different companies, think Ribbon, Katana, Friction, and a few others who built covered call vaults on blockchains like Ethereum, Solana, and others. The idea there was you could deposit SOL, ETH, or Bitcoin, you could sell calls against it and collect the premium. And actually, that product's worked in TradFi forever. The problem fundamentally is that SOL or ETH or, you know, any of these crypto assets, they can go up or down 50% in a month. And so if you're giving all of that upside away for a few points of premium, the product just is not that compelling. But if you point that same exact product at an equity that, you know, say might appreciate 10% in a year, a covered call actually starts to make a lot of sense. And so do other interesting structured products, things like collars and downside protection products. So we think that option vaults weren't necessarily wrong. Again, they were just built for the wrong assets at the right time.

9:07其實整個加密選擇權市場也是如此。永續合約會在加密圈勝出，是因為標的資產波動率高達 100%，你開 3 倍多單，即使是 delta one 衍生品，也能玩出很多花樣。但當標的變成國庫券或股票時，鏈上選擇權就會像傳統金融一樣重要。And honestly, I think that's true of the entire options market and crypto more broadly. You know, perps have won in crypto because a lot of the underlyings have 100% volatility. So if you're 3x long, I don't know, a hype perp, even though it's a delta one derivative, you can still get a lot of skew and, you know, maybe some dopamine from that. But once the underlyings or equities, I expect options on-chain to matter just as much as they do in traditional finance.

9:28第二個差異是存續期間。Bitcoin、ETH、SOL 等加密資產沒有到期日，但現實世界的金融契約幾乎都有。例如公司借 1 億美元三年做資本支出，必須知道借款成本是每年 500 萬還是 1,000 萬；資產管理人買 6% 殖利率的債券，用 4% 成本融資，中間 2% 就是利差。如果隔天融資利率變成 6%，這筆交易就沒意義了。在這些情境下，利率確定性從「加分」變成「必要」。所以固定利率、固定期限的信貸產品會變得很重要。像 Morpho 的 Midnight，就是用特定到期日的買賣報價來決定利率，這樣就能建立真正的殖利率曲線，三個月、六個月、一年，甚至更長。這對產業很重要，因為所有東西都可以用這個基準定價，借款人也能真的拿到三年期資金，殖利率曲線本身還能反映市場對利率和風險的預期。DeFi 現在根本沒有這種訊號。The second difference between traditional assets, i.e. RWAs, and crypto assets is duration. Bitcoin, ETH, SOL, etc., they obviously don't have a maturity date, but most real-world financial obligations do, right? They have dates attached to them. So if a company is borrowing $100 million for three years to, I don't know, fund some CapEx project. It probably needs to know whether its borrowing cost is going to be $5 million a year or $10 million, or if an asset manager buys a bond that's yielding 6% and they're borrowing at 4%, the 2% spread is the trade. If funding moves on them to 6% tomorrow, the trade's obviously gone. In these cases, in the real world, rate certainty goes from, you know, a nice to have to basically essential. So that's why fixed rate, fixed term credit becomes so much more useful. Morpho, with their new Midnight product, for example, is doing this today with rates that are set by bids and offers at specific maturity dates. And then once you have those markets, you can actually start to get a real yield curve like you see on the slide. So things like three-month yield, six months, one year, eventually even multi-year. We think that this is generally pretty important for the industry because it's a benchmark that everything else can get priced off of. A borrower with a three-year obligation can actually get funded for three years, and also the curve itself can tell you what the market expects for rates and risk over time. DeFi today essentially does not have that signal.

10:59現金流、利率衍生品與期貨新契機

10:59另一個有趣的地方是，當你有了殖利率曲線，利率本身也能被交易。像 Pendle 和我們投資的 Exponent，就把本金和利息分開交易。以前 DeFi 上大家交易的殖利率是質押收益、Aave 借貸利率、Binance 的資金費率，甚至是 DeFi 協議的積分。但有了 RWA，大家就能交易國庫券利率、信貸利差、股息，甚至任何合約現金流。Boros 也做了類似的固定與浮動利率交換，現在主要用在中心化交易所的永續合約資金費率，但未來這種結構可以幫你避險真實的利率風險。And then another interesting thing is once you actually now a curve rates themselves become tradable. So Pendle and our portfolio company Exponent, for example, they separate out principal from yield. Historically, the yields that people were trading in DeFi were things like the staking yield or the borrow rate on Aave or funding rates on Binance or eventually actually points on DeFi protocols. With RWAs, though, it becomes things like treasury rates, credit spreads, dividends, and really any sort of contractual cash flow that you can imagine. Boros, which is a different Pendle product, actually does the same for fixed versus floating swaps. Today it's just on perp funding on the major centralized exchanges, but in the future that structure could actually help you hedge out some of these real interest rate risks.

11:49最後，傳統有到期日的期貨合約也會在加密圈變得重要。過去沒人用，是因為標的資產沒有存續期間。但如果我六個月後要付歐元，曝險就有明確的到期日，避險工具也應該要有。And then lastly, dated futures will actually start matter in crypto. I know that's kind of a taboo thing to say. Historically, nobody's really used them, and that's because the assets that you're trading don't have duration. But if I owe someone euros in six months, my exposure has a date, and my hedge probably should as well.

12:08RWA 抵押品、借款人與機構參與

12:08第三個差異是抵押品價值。RWA 比加密資產好太多。早期的鏈上借貸或穩定幣協議，都假設抵押品可能一夜間跌三成、五成，所以像 MakerDAO 早期要你拿 150 美元 ETH 才能借 100 美元 DAI。但如果你用國庫券抵押，借款人還是匿名，折價可能只要 10 美元，LTV 可以到 91%。如果是已經 KYC 的機構用國庫券做抵押，甚至能借到 98%。更好的抵押品已經解決大半問題，剩下的就是你能不能知道對手方是誰。The third difference between RWAs and traditional crypto assets is just the value of their collateral, right? RWAs vis-a-vis are just much better collateral than traditional volatile crypto assets. In early crypto lending protocols or even stablecoin protocols, you kind of had to assume that the collateral could drop a third, 50% overnight, and that's why MakerDAO in the early days had you put up $150 of ETH to borrow out $100 of DAI. But if you swap out ETH for treasuries and you keep the borrower anonymous, maybe that haircut falls to $10, right? So now you're looking at a 91% LTV instead of 66. And then if you look at the third row on the chart here on the left, if it's a known institution who's KYC'd on-chain that's posting treasuries in the repo market, maybe they can borrow at 98%. So better collateral will get you a lot of the way there. And then the last step, which I'll talk about a little bit on the next slide, is actually knowing who's on the other side of the loan.

13:05RWA 也改變了借款人輪廓，因為你現在可以真正審核對方。Ethereum 錢包沒有財報，但公司有。你可以看現金流、負債、管理團隊、還款紀錄，最重要的是如果違約你有什麼法律追索權。Maple 已經在鏈上做機構借貸，早期 DeFi 的無抵押借貸網路很難做，因為逆向選擇嚴重，來借錢的往往是其他地方借不到的。隨著美國監管明朗，合法企業會開始上鏈借款，這會為全球放款人創造很多風險調整後的好機會。So like I just mentioned, RWAs can change the borrower profile too, because now you can actually underwrite them, right? An Ethereum wallet, it doesn't have financial statements, but a company obviously does. So you can look at things like cash flow and liabilities, the management team, their repayment history, and most importantly, what legal claims you actually have if there's a default, right? If I go bust on a decentralized derivative exchange, no one can do anything about it. Maple is already doing a version of this on-chain with institutional borrowers, and candidly, a lot of the early undercollateralized lending networks in DeFi struggled because there was just a ton of adverse selection, right? The borrowers who showed up to these things like a Goldfinch were, in a lot of cases, the ones who couldn't really get credit anywhere else, either because of just their creditworthiness or maybe the jurisdiction they were operating in. I think, though, you know, as the regulatory clarity improves in the U.S., legitimate businesses can actually start to borrow on-chain, and that will create a lot of interesting risk-adjusted opportunities for lenders globally.

14:11機構參與與鏈上投資組合的未來

14:11第四個差異是誰在交易，以及單筆訂單的規模。如果我在鏈上買 1 萬美元 ETH，沒人在乎透明度。但如果我是資產管理人，要分批買 1 億美元的中型股，把訂單攤在市場上會非常昂貴。我們 Multicoin 自己就遇過這種事，當我們在基金裡建倉 HYPE，任何有錢包追蹤工具的人都能即時看到我們的部位。我們試著把代幣轉到新錢包想「去標籤」，結果市場以為我們要賣，大家都恐慌。我們只好出來澄清。傳統金融早就用 block trade、RFQ、dark pool 等工具解決這個問題，這些結構現在也開始被搬到鏈上。我預期這些協議很快會佔據鏈上交易量的重要份額。The fourth difference here between RWAs and crypto assets is who's actually trading them in a lot of cases, and then, like, how big those orders are. So if I buy, you know, $10,000 worth of Ether on-chain, nobody really cares that it's transparent and visible. But if I'm an asset manager that's doing a TWAP for $100 million of some mid-cap equity, showing that order to the market can be extremely expensive. And we at Multicoin actually learned this one firsthand. So when we were building a position in Hype in our hedge fund, anyone with a wallet tracker could see our position in real time. So we then tried to go and un-Dox ourselves by moving the tokens to new wallets. And when we did so, the market absolutely freaked out because people thought we were selling, and we had to tell everyone that we weren't. So TradFi solved this a long time ago with primitives like block trades. It's RFQ systems and dark pools, and those structures are now starting to come on-chain by some of the companies that you see in the bottom left here. I expect a lot of these protocols to be a meaningful share of on-chain trading volumes within just a few years.

15:16最後一個、但我認為最被低估的差異，是投資組合本身的樣貌。以我自己為例，我住在美國，可能會在 Coinbase 持有 Bitcoin、E-Trade 持有股票、Robinhood 買選擇權、Kalshi 玩預測市場、CFD 平台做外匯、Clarity 這種平台持有私募股權。這代表我有六個帳戶、六套保證金制度、六池抵押品，彼此完全不互通。舉例來說，如果我在一個帳戶做多股票，在另一個帳戶買了保護性賣權，沒人會看到這個避險組合，我還得重複抵押。其實現在 DeFi 也沒好到哪裡去，因為大家持有的資產本來就都是同一種 beta。但當利率、信貸、股票、外匯等資產都上鏈後，真正的資產組合抵消效應就會出現。我們認為 DeFi 原生的 prime brokerage（主經紀商）產品會非常有趣，我們投資的 Project Zero 就在 Solana 上做這件事。終極願景是你能用一個資產負債表管理所有資產類別，實現真正的投資組合保證金和資本效率最大化。So the last difference I'll highlight, which I think is actually one of the most underappreciated, is what the portfolio itself looks like. If you think about how, you know, I live in the U.S., someone like me would hold assets today. I might hold Bitcoin on Coinbase, equities on E-Trade, options on Robinhood, prediction markets on Kalshi, FX on some sketchy CFD broker, and then maybe private equity shares on a platform like Clarity. So I would have six different accounts with six different margin systems and six pools of collateral, and none of them would actually talk to each other. So as an example, if I was long an equity in one account, but I owned a put option in a different one, nobody would see the hedge and I wouldn't get credit for it, so I'd have to post collateral twice. Honestly, the way DeFi works today isn't much better. In crypto, though, that hasn't really mattered because most of the assets that people own are just the same beta anyways. But once you have rates and credit and equities and FX and all these other assets we're talking about coming on chain, there's actually going to be real offsets everywhere that you can net against each other. And so we think that a DeFi native prime brokerage product is extremely interesting, and our portfolio company Project Zero is actually building on Solana right now. The grand vision here is you have one balance sheet across every asset class that is entirely portfolio margined and allows you to get maximum capital efficiency.

16:42DeFi 2.0 的投資機會與產業擴張

16:42那 Multicoin 為什麼會關心這一切？我認為這裡有兩個市場規模（TAM）擴張同時發生。第一個很直觀，就是資產種類變多。全球債券約 161 兆美元，股票約 158 兆美元，而整個加密市場不到 3 兆。鏈上 RWA 目前大約 390 億美元，只佔全球股票和債券的萬分之一。就算只把傳統金融的一小部分資產搬上鏈，DeFi 的服務範圍也會大幅擴大，現在還只是剛起步而已。So the natural question is, why do we at Multicoin actually care about any of this from an investment perspective? I think there's really two different TAM expansions that are happening at the same time. The first one is obvious, but I think the second is actually pretty interesting. So the first obvious one is simply more assets, right? Global bonds are about 161 trillion, equities are about 158 trillion, and all of crypto. I don't even looked at prices today, but call it under three trillion. Tokenized RWAs are about 39 billion, so roughly one basis point, one hundredth of 1% of global stocks and bonds. So even moving a tiny fraction of TradFi assets on-chain is an enormous expansion of what DeFi can serve, and we're obviously very early here.

17:30第二個 TAM 擴張更有趣：RWA 每一美元都需要更多金融基礎設施。想像一個 Solana meme coin，頂多需要現貨交易、永續合約，或有一點抵押價值就能有借貸市場，需求就這樣而已。但國庫券就不一樣了，你可以做現貨交易、repo 融資、抵押借貸、期貨避險、利率衍生品，還能包裝成結構型商品。換句話說，不只是資產變多，每一美元都能創造更多金融活動。The second TAM expansion, though, I think is arguably the more interesting one, which is that RWAs need a lot more financial infrastructure around every dollar. So think about a meme coin on Solana, right? Maybe it needs spot trading, maybe a perp if there's enough liquidity, and then maybe a lending market if it has any collateral value whatsoever and/or there's enough demand to short it, and that's basically it. A treasury is different. So you can trade it spot, you can finance it through repo, you can post it as collateral, you can hedge it with futures, you can trade the rate separately via interest rate derivatives, and you can package it into compelling structured products that, you know, you would never do with a meme coin. So we're not just putting more assets on chain, we're actually creating a lot more financial activity around every dollar that comes on chain.

18:20誰會在 DeFi 2.0 賺到錢？

18:20那誰會從這一切賺到錢？我認為整個堆疊有四個層級的機會。最底層是區塊空間，交易、融資、結算需求增加，L1、L2 都會受惠。再來是核心 DeFi 模組，像交易所、借貸、選擇權、利率衍生品等協議。第三層是聚合層，也就是剛剛提到的主經紀商。最上層則是擁有使用者關係的應用端。我們在部落格寫過 publisher exchange thesis，意思是擁有使用者的 app 決定資金流向哪個協議、抵押品放哪裡、收益怎麼分配，這讓他們能從底層協議攫取大量利潤。通常你在堆疊越上層，能抓到的利潤越多，但現在還很早，網路效應會落在哪裡還不確定。The obvious question then, you know, I'm a VC, is who makes money from all of this? I think there are generally four opportunities kind of up and down the stack. So at the bottom is block space. So obviously more trading, more financing, and more settlement generally means a persistent source of demand for chains, whether it be an L1 or an L2. And then next, the core DeFi primitives themselves are going to make money, things like exchanges, lending protocols, options protocols, rate derivatives protocols, these types of things. And then aggregation, which is the prime brokerage layer we just talked about. And then at the top, you know, whoever owns the customer relationship. We've written about this actually on our blog. We called it, I believe, the publisher exchange thesis, which is basically that the app that owns the customer decides where trades get liquidity from, where the collateral sits, where yield is routed and which product the user sees, and this gives them an enormous amount of power from all of the protocols that are underneath them. I think broadly, the higher up the stack you are, the more margin you're able to capture. But again, it's obviously very early, and we don't know where the network effects are actually going to accrue just yet.

19:30DeFi 2.0 不是重頭來過，而是資產終於跟上

19:30這張投影片很重要。DeFi 不需要從零發明這些新模組，因為很多早就做出來了。Ribbon 幾年前就做了 covered call vault，Friktion、Katana 也有，但標的是高波動加密資產。Exponent 把本金和殖利率分開，TVL 也不錯，但大家交易的殖利率主要還是積分。Boros 做了利率衍生品，但最熱門的標的是永續合約資金費率。Project Zero 做了統一保證金協議，但現在還只是淨額化加密曝險。我們的看法是，這些產品不是錯，只是太早，資產還沒準備好，但現在終於要來了。This slide, I think, is quite important. DeFi doesn't need to invent these new primitives that I'm talking about from scratch because we've already built a lot of them, right? Ribbon built covered call vaults years ago, and then so did Friction and Katana. But again, the underlying was volatile crypto. Exponent, they separate principal from yield, and it's actually scaled pretty nicely in terms of TVL. But the most interesting yield that people are trading right now has been point programs, which is just like not that compelling. And then Boros built rate derivatives, but again, the product that people trade the most is perp funding. And Project Zero, which I mentioned earlier, has built a unified margin protocol, but again, it's just really netting crypto exposure. So our view is that these products are not wrong, they're just early, and the assets have not been ready yet, but they're now coming.

20:17我要強調，DeFi 1.0 不會消失。AMM 會繼續服務長尾代幣，永續合約還是加密投機的主力，浮動利率借貸也適合想加槓桿的匿名使用者。但我想總結的是，RWA 第一階段是把資產搬上鏈，這也是大家現在努力的方向。下一階段則是圍繞這些資產建立真正的金融市場。DeFi 的基礎設施早就準備好了，現在資產終於跟上，這就是我們所說的 DeFi 2.0。So, to be clear, I don't think DeFi 1.0 is going to go anywhere, right? AMMs will continue to serve the long tail of tokens. Perps will remain a main tool for crypto speculation, and variable rate lending still makes sense for borrowers who are pseudonymous and just want to lever up on their crypto assets. But the way I'd summarize the talk is, you know, the first phase of RWAs has been about getting assets on-chain, and that's what a lot of people here are working on. But the next phase is about actually building real financial markets around these assets. In DeFi, we've already built a lot of these primitives, and so the assets are now finally catching up, and that's what we mean by DeFi 2.0.

20:55結語與行動呼籲

20:55今天內容很多，如果你想看完整的 DeFi 2.0 論文，可以掃這裡的 QR code。若你是 builder，正在解決這些問題，歡迎會後找我聊，或在 Twitter 上聯絡我（@Spencer Applebaum）。謝謝大家！So I know I covered a lot of ground today. If you want to read the full essay on our website that kind of unpacks our DeFi 2.0 thesis, you can scan the QR code here. If you're a builder that are working on any of these problem sets, please come find me after the talk or reach out to me on Twitter at Spencer Applebaum. Thank you.