機構要自己的鏈，為什麼選二層而不是一層？

0:00機構要自己的鏈

0:00今天要特別談 Arbitrum，以及為什麼我認為可程式經濟（也就是機構資金正在上鏈）會發生在二層，而且會發生在 Arbitrum。問題已經不是金融會不會上鏈，這正在快速發生。問題是在哪一條生態？我們看到機構多半想自己做一條鏈。這常常是起點。最近 Robinhood 做了 Robinhood Chain，Coinbase 做了 Base。Today I'm going to talk particularly about Arbitrum and why I believe the programmable economy, which is the influx of institutional finance that we're finding happening on-chain today, is going to happen, number one, on Layer 2, and number two, on Arbitrum specifically. The question today isn't whether finance is going to move on-chain. We're seeing that happen rapidly. The question, though, is where? Which ecosystem is this going to happen in? And we're seeing that institutions predominantly want to build their own chains. And this is often a starting point for institutions. If you look recently at Robinhood, built Robinhood Chain; Coinbase built Base.

0:46Circle 做了 Arc，Stripe 做了 Tempo。這些機構通常面對兩個選擇：做在以太坊二層，或另做一條一層。最大的那些機構今天在問的就是：做在以太坊，還是自己做一層？對以太坊來說，這位置其實很好，因為其他鏈被排除了。他們不是在問要不要在別條鏈上做二層，而是：做在以太坊，還是自己做。Vitalik 2020 年那篇 rollup 為中心的路線圖其實就寫到了。很多人以為 rollup 只為了擴容，但你可以去讀原文，我打在螢幕上：另一個原因是，人會想當自己小島上的王。Circle built Arc; Stripe built Tempo; and these institutions are typically faced with one of two choices. Either they're going to build on Ethereum Layer 2, or they're going to build an alternative Layer 1. Those are the choices that the largest institutions are having today. Do I build on Ethereum, or do I build my own Layer 1? And by the way, that's an incredible place for Ethereum to be, because it basically knocks out every other blockchain. They're not saying, hey, do I want to build a Layer 2 on some other chain? It's, do I want to build on Ethereum, or do I want to build my own thing? And by the way, Vitalik actually called this exactly in his 2020 post. In 2020, Vitalik published the well-known rollup-centric roadmap for Ethereum, in which he basically said, you know, people get this wrong a lot. People often think that rollups are only for scaling, but what Vitalik said, and you can read the post, I have it up on screen here, is another reason that people are going to want to build on Layer 2 is because people are going to want to be the king of their own little island, right?

1:45機構不想在大型公共網路上當小角色，而常想在自己的小島上當大玩家。這點很多人沒搞懂。今天常有人把這段歷史講錯，說以太坊該把所有東西做在一層，有了擴容就不需要二層。但二層原本的動機之一，正好就是我們現在看到的：全世界最大的那些機構想要自己的鏈、自己的小島。所以若你是這種機構，選擇就是：自己做一層，還是做以太坊二層。接下來拆開看利弊，以及為什麼我認為大型機構今天選二層。Institutions are not going to want to be a little player on a large public network, but they're often going to want to be a large player on their own little island. And that's something that's not well understood. Today, people often misquote this history, and they say, hey, look, Ethereum should just build everything on Layer 1. We don't need L2s anymore. We have scale. But again, part of the original motivation for Layer 2s was actually exactly what we are seeing today from the world's largest institutions, that they want to have their own chains. They want to have their own little island. So if you're one of these institutions, you basically have one of two choices: do I build my own Layer 1, or do I build an Ethereum Layer 2? So let's break that down. Let's see what are the pros and cons of building a Layer 1 or Layer 2, and why I believe the large institutions are choosing a Layer 2 today.

2:31一層很貴，二層留下約 97%

2:31第一個問題是經濟。大家不太愛談，但一層在經濟面有麻煩：營運很貴，啟動也很貴。若要認真做一層，有分散的驗證者集合、真正的安全性，你就得為安全性付錢，成本往往非常高。任何一層，所有費用都會直接轉給驗證者。一層自己留不住經濟活動帶來的錢，得全數轉出去。這跟你有沒有成功無關。鏈上只要有資產，就得付安全性。So the first question is economics. People don't like to talk about this, but L1s have a bit of an issue on the economic side, and that is they're really, really expensive to run and really expensive to bootstrap. If you want to properly build a Layer 1 with a distributed validator set and real security, you have to pay for that security, and that often has a very, very high cost. In fact, all of the fees, if you take any Layer 1, all the fees from that network get passed right through to their validators. The Layer 1s retain zero dollars of their economic activity, and they have to pass those dollars through to their validators. And that's true, by the way, if you're an institution thinking, hey, do I want to launch a Layer 1 or Layer 2? That's true whether or not you have any success. If you're running a blockchain that has assets on it, you need to pay for its security.

3:16一小時只有一筆交易，安全性還是要付，成本常常是一年數億美元，只為了讓鏈活著，跟商業成不成功無關。有人把這件事講反了，說一層代幣很好，因為有孳息。那孳息其實就是安全性的成本，是通膨的成本。一些最受歡迎的替代一層，為了付安全性，代幣供給每年通膨超過 5%。交易費每一美元都給驗證者，還不夠付驗證，於是再靠通膨補安全性。有人說「我質押有孳息」；其實你只是保住比例、沒被通膨燒掉，因為跑一層太貴了。If there's only one transaction an hour, you still need to pay for the security, and that cost is often in the hundreds of millions of dollars per year just to keep the chain up, regardless of whether you have any commercial success. And people often flip this on its head. They say, "Oh, but Layer 1 tokens are great because they give me yield." That yield is actually just the cost of security. It's the cost of inflation. If you look at some of the most popular alternative Layer 1s, they're inflating their token supply year over year over 5% just to pay for the security of the chain. What's happening is every dollar of these fees generated is going to the validators, but that's not enough to actually pay for validation. So instead, they're inflating their token supply to pay for the cost of security. And people sometimes get this wrong. They say, "Oh no, no, I'm getting yield. I'm staking, getting yield." No, your "getting yield" is you just basically keeping your pro rata share and not getting burned by the inflation, because the cost to run a Layer 1 is so, so expensive.

4:11二層結構不一樣。二層的安全性成本是每筆交易很低的固定成本。Arbitrum、Robinhood Chain 和其他以太坊二層，大約把 3% 到 5% 的交易成本付給以太坊。就這樣。這條鏈大約有 97% 的利潤率。一層世界裡，交易費全給驗證者，還得通膨付安全性；二層不必通膨，交易費大約 97% 留在協議。所以 Arbitrum 和其他二層能累積金庫、留下成長、把成長再投入生態。Layer 2s, though, are built differently. If you're a Layer 2, you know what your cost of security is? You pay a very low fixed cost that's per transaction. Arbitrum and Robinhood Chain and other Ethereum Layer 2s pay about 3 to 5% of their transaction costs to Ethereum. That's it. The chain is run with 97% margin. So whereas in Layer 1 world, all transaction fees go to the validators, and then you have to inflate the token supply to pay for security. In Layer 2 world, actually, you don't have to inflate at all, but actually, even the transaction fees, 97 or so percent are retained by the protocol. And that's why you have Arbitrum and other Layer 2s, the ability for them to build treasuries. And actually retain their growth and compound their growth and invest in other ecosystem initiatives.

4:58Arbitrum 在二層裡特別的是：多數二層把交易費導到基金會或某個不夠透明的實體；Arbitrum 的交易費直接進鏈上、由代幣持有人控制的錢包。一層留不住收入，還得通膨；二層可以留下約 97% 的收入。所以機構若還不確定這條鏈會不會成功，二層比較安全：費用上限只是成功的幾個百分點。交易多，大約付 3%；交易為零，安全性跟著縮，保全這條鏈幾乎不用錢。And Arbitrum, in particular among Layer 2s, is unique because most Layer 2s that have this, they funnel those transaction fees to some foundation or some entity that doesn't have exact transparency. In Arbitrum, all the transaction fees actually go directly on-chain to the token holder-controlled wallet. So Layer 1s don't have the ability to retain revenue. Layer 1s need to inflate their token supply, whereas Layer 2s, you can actually retain 97% of your revenue. So if you're an institution today thinking, do I build a Layer 1 or Layer 2, and you're not sure how successful your chain will be, a Layer 2 is a much safer choice because your fees are capped at just a few percentage points of your success. So if you have a lot of transactions, you'll pay about 3% of that. If you have zero transactions, the chain security will scale down with you, and it'll still cost you almost nothing to secure your chain.

5:47分層設計，不必重做整疊

5:47一層則不論一天只有一筆交易，往往仍得一年付數千萬到數億美元，只為了讓鏈活著。第二個想談的，是一層對二層的技術架構。分層其實合理。史上最成功的協定，你們可能聽過，叫網際網路，就是分層做的。想像每做一個網站都得重做整套網路堆疊，不能用 TCP 或 UDP，顯然很荒謬。分層的好處是：你做應用，底下那些自動接上，不必自己煩。Whereas on a Layer 1, you'll actually have to, no matter how many transactions you have, if you have one transaction per day, you need to continue to pay often tens to hundreds of millions of dollars a year just to keep the chain alive. A second area where I'd like to focus on the Layer 1 versus Layer 2 debate is in the architecture, the technical architecture. And a layered design actually just makes sense. The most successful protocol of all time, you may have heard of it, it's called the Internet, is actually built in layers. And if you imagine building a website, and every time you want to build a website, you actually have to rebuild the entire network stack. So you can't tap into TCP or UDP. You actually have to rebuild the entire network stack. Obviously, that's ridiculous. One of the nice things about the layered internet stack is that you can just build an application and automatically tap into everything underneath you and not have to worry about it.

6:39一層對二層也一樣。做二層，你可以在鏈這層創新，不必自己啟動共識、找驗證者、付驗證者，那些交給以太坊。就像你不會為了做網站去重做整套網路堆疊。2026 年已經有 Arbitrum 這種建在以太坊這種穩固一層上的二層，每次再重做那一疊，沒有道理。所以不論經濟還是技術，對多數今天的機構，二層就是比較好的設計。我很堅持（你可以來說服我）：今天幾乎任何想自己做鏈的人，二層都好很多。It's the exact same thing for Layer 1 versus Layer 2. If you're building a Layer 2, you can innovate at the chain layer. You don't have to worry about bootstrapping your own consensus. Don't have to worry about finding your own validators. Don't have to worry about paying your own validators. You can use Ethereum for that. Just like you wouldn't rebuild your entire network stack just to create an application-layer website. Similarly, it doesn't make sense in 2026, when we have L2s like Arbitrum built on strong and secure L1s like Ethereum, to rebuild that stack every time. And that's why, both from an economic perspective and a technical perspective, Layer 2s are just a better design for most institutions today. So I believe very firmly—convince me that I'm wrong, change my mind—that Layer 2s are the much, much better choice for just about any builder today that wants to build their own chain.

7:28為什麼是 Arbitrum，以及 USDG

7:28若要做二層，Arbitrum 特別突出，是最獨特的二層生態，接下來幾分鐘談這個。大家對 Arbitrum 知道什麼？技術口碑好，區塊時間是加密貨幣裡最快的，手續費極低，吞吐量高。Robinhood Chain 建在 Arbitrum 平台上，最近創下交易速度和交易量紀錄，超過任何 EVM 鏈做過的。除了技術，還有深的流動性：你可以立刻把 DeFi 應用發到 Arbitrum 活躍的 DeFi 生態，市場活、資產也多。And in particular, if you're looking to build a Layer 2, Arbitrum stands out and is the best and most unique Layer 2 ecosystem to build in, and that's what I will focus on for the next few minutes. So what do we all know about Arbitrum? Arbitrum has a reputation for having excellent technology. It has the fastest block times in all of crypto. It has extremely low transaction fees. It has high throughput. We recently saw Robinhood Chain built on the Arbitrum platform, doing record transaction speeds, record transaction volume, more than any EVM chain has ever done. Arbitrum also has, besides its technological advances, deep liquidity. You can immediately launch a DeFi app into Arbitrum's thriving DeFi ecosystem, with deep liquidity and vibrant markets and many, many assets also.

8:14整個生態裡，Arbitrum 有超過 7,000 筆 RWA，比加密貨幣裡任何其他生態都多。所以在二層裡，尤其是看重 DeFi 和交易的，Arbitrum 是很好的地方。Offchain Labs 和 Arbitrum 生態對齊的一件事是：把鏈的技術和生態設計成讓建造者贏。所以區塊時間短，適合發行資產、也適合交易這些資產、讓資產進入會競爭的狀態。我們在生態裡做的每件事，都圍繞交易、金融和競爭狀態。Across the ecosystem, in fact, Arbitrum, with 7,000-plus RWAs, has more RWAs than any other ecosystem in all of crypto. So Arbitrum is a fantastic place, particularly among Layer 2s, one of the most vibrant Layer 2s, with its focus on DeFi and trading, to build your application. And in particular, we at Offchain Labs and we in the Arbitrum ecosystem are aligned on one thing. We want to architect our chain. We want to architect both the technology and the ecosystem to make Arbitrum builders win. And that's why you have low block times that are optimized, for example, for both launching assets, but also trading those assets, for having those assets enter contended state. And everything that we do in the Arbitrum ecosystem is focused on trading, finance, and contended state.

9:01另一個例子：我們昨天宣布與 Paxos 的 Global Dollar Network（USDG）合作。這很重要，因為又給 Arbitrum 建造者多一條贏的路。Arbitrum 生命週期裡，創造了數十億、數百億美元的穩定幣需求和成交。但直到昨天，生態系從中拿到的是整整零美元。平台上有這些穩定幣、有這些量，孳息卻被第三方拿走。光 Arbitrum 上的穩定幣，一年孳息就是數億美元，沒回流生態。昨天開始變了，Arbitrum 加入了 USDG。Another great example: we just announced yesterday a historic partnership with USDG, the Global Dollar Network from Paxos. And the reason why this is so important is it gives Arbitrum builders yet another way to win. So Arbitrum has, over the course of its lifetime, generated billions and tens and hundreds of billions of dollars in stablecoin demand and stablecoin volume. But up until yesterday, the ecosystem captured exactly zero dollars of that. So Arbitrum has all these stablecoins on the platform, is generating all this volume, but other third parties are actually taking all the yield. And, you know, all in, it's hundreds of millions of dollars a year of yield generated from Arbitrum stablecoins alone that's not being actually captured and not benefiting the ecosystem. That changed yesterday, and from now on, Arbitrum actually has joined USDG.

9:49Arbitrum 現在有自己的原生穩定幣布局。在 Arbitrum 上，USDG 穩定幣的回饋會回到生態、回到代幣持有人控制的金庫、回到 Arbitrum 協議。這是協議把成功再疊上去的另一條路。機會很大，因為 Arbitrum 上傳統金融和 DeFi 協議都很多。加密貨幣很愛加形容詞：TradFi、DeFi。Arbitrum 兩邊都有。但拉開看，不是兩個圈，是一個正在會合的圈，那圈不叫 DeFi，叫金融，而且發生在 Arbitrum。Arbitrum now has its own native stablecoin play. So if you are on Arbitrum, the rewards from the USDG stablecoin on Arbitrum are going back to the Arbitrum ecosystem, back to the Arbitrum token-holder-controlled treasury, and back to Arbitrum protocols. So, yet another way for Arbitrum protocols to compound their success. This will have a massive opportunity because Arbitrum has so many protocols that are thriving across TradFi and DeFi. You know, I like to say, I've said this for a very long time, although it's becoming a bit more popular now, we like our adjectives a lot in crypto. We have TradFi, we have DeFi. Here we see Arbitrum has both of them. But if you zoom out here, and you really, really zoom out, you don't see two circles. There's really just one circle that's converging, and that circle isn't called DeFi. That circle is called finance, and that's happening on Arbitrum.

10:36不是每個人都要自己的鏈

10:36無論 DeFi 還是傳統金融，Arbitrum 上的金融協議在一起做事，更有機會拉使用者、創造孳息，讓建造者和使用者都贏。幾個在生態裡做得好的例子。上面是 Variational，原生於 Arbitrum、較新的永續合約協議。永續合約這條線，Arbitrum 在它變熱門之前就在做，從 GMX、Gains Network 開始。Hyperliquid 歷史上流動性來自 Arbitrum；Lighter 一開始也在 Arbitrum 上線。今天 Ostium、Variational 都在這條網路上；DeFi 和永續合約從第一天就是 Arbitrum 的基因。And Arbitrum protocols, financial protocols, whether they're DeFi or TradFi, are working together, have more opportunities to both attract users, create yield, and, you know, to have builders and users alike win in the Arbitrum ecosystem. Here are some examples of some of the great businesses that are doing well in the Arbitrum ecosystem. On the top here, we have Variational. Variational is the newest and up-and-coming perp protocol that's native to Arbitrum. If you look at perps history, you know, Arbitrum was doing perps well before they were cool. Of course, starting with GMX and Gains Network. Hyperliquid, of course, historically sourced all their liquidity from Arbitrum. Lighter initially launched on Arbitrum. Today we have protocols like Ostium and Variational that are absolutely killing it on the Arbitrum network, and DeFi and perps really have been native to the Arbitrum DNA from day one, and that continues to be what makes Arbitrum stand apart.

11:26Spiko 在 Arbitrum 上的量也很大。Bitget Wallet 剛在 Arbitrum 上推出代幣化股票，和 Robinhood 的股票並列，讓 Arbitrum 成為代幣化股票、代幣化股權和 RWA 的主要場地。USD.AI 是另一個在 Arbitrum 上做得好的協議。重點是：好生意用 Arbitrum。我們會繼續加像 USDG 原生孳息分享這種會疊加的機會，讓 Arbitrum 在 2026 年以後仍是做生意最好的地方。到這裡我說的是：若你想像 Coinbase、Robinhood、Circle、Stripe 那樣自己做鏈，最好的決定是做二層。Spiko, of course, doing incredible, incredible volumes on Arbitrum. You have Bitget Wallet, which just launched tokenized stocks alongside Robinhood stocks on the Arbitrum platform, making Arbitrum the leading player for tokenized equities, tokenized stocks, and RWAs generally. And USD.AI is another fantastic example of an innovative protocol that's killing it on Arbitrum. And the takeaway here is that great businesses use Arbitrum, and we in the Arbitrum ecosystem continue to add more and more compounding opportunities, such as native yield sharing with USDG, to make sure that Arbitrum is the absolute best place to build a business. In 2026 and beyond. So I've focused until now on saying, if you want to build your own chain, which increasingly institutions like Coinbase, institutions like Robinhood, institutions like Circle, institutions like Stripe are doing today, the absolute best decision you can make is to build a Layer 2.

12:17在二層裡，最好的決定是做在 Arbitrum。但不是每個人都需要自己的鏈。這點 Arbitrum 特別：我們可以說，不必人人都做自己的鏈。競爭對手通常分成兩派。一派說每個人都要自己的鏈，不論你是誰、規模多大。另一派說公共網路就好，沒人需要自己的鏈，那些鏈很荒謬。真正的答案在中間、比較細。像 Robinhood Chain 這種較成功的協議，很吃自己的鏈；生命週期較早的 DeFi 協議，則受益於公共網路的公共流動性，能很快起步。And within the Layer 2 space, the absolute best decision you can make is to build on Arbitrum. But does everyone need their own chain? And this is a place where Arbitrum actually stands out, where we can say, no, not everyone needs their own chain. My competitors typically fall into one of two buckets. There are those that say everybody needs their own chain. Everyone. Doesn't matter who you are, doesn't matter what scale you are, everyone needs to launch their own blockchain. And there are those that say, no, no, no, public networks. No one needs their own chain. These chains are ridiculous. Everyone should just launch on the public network. But the real answer is actually somewhat nuanced in between that. Some of the more successful protocols, like Robinhood Chain, significantly benefit from their own chain, but other protocols, including DeFi protocols earlier in their lifecycle, actually benefit from being on public networks with public liquidity, where you can get started off very quickly.

13:04問題是，別的網路、別的鏈堆疊，會在第一天就逼你選：要自己的鏈，還是上公共網路，兩條路從此分開。Arbitrum 兩邊都有，路是會合的。你可以先在 Arbitrum One 起步，那是今天做 DeFi 流動性最好的地方之一；到對的時間、對的成功點，再遷到自己的 Arbitrum 鏈。不必第一天就決定。大家都以為自己會走到需要自己的鏈；有人會，有人不會。Arbitrum 可以陪你走完整段生命週期，這點是獨有的。沒有其他生態同時有 Arbitrum One 這種中立的頂層公共鏈，又有頂層的發鏈堆疊，還能在兩者之間平滑移動。The problem is, if you look at other networks and, you know, our competitors both on the chain side and the chain-stack side, they're going to push you into that decision on day zero. You have to decide today at launch, do I want to optimize for launching my own chain, or do I want to launch on a public network? And those paths diverge. But in Arbitrum, it's unique because we have both of those paths. Those paths actually converge. You can start off on Arbitrum One, one of the most liquid and best places to build a DeFi protocol today. And at the right time, at the right point of success, you can migrate to your own Arbitrum chain. You don't have to make that decision on day zero. Everyone thinks they're going to get to the point where they need their own chain. Some will, some won't. On Arbitrum, we can serve you across the lifecycle, and that's unique. There is no other ecosystem out there that has a top— incredibly neutral, top blockchain like Arbitrum One, but also the top blockchain stack to launch your own chain, and you can seamlessly move between them.

13:58Robinhood 先上 One，再搬自己的鏈

13:58這是有先例的。Robinhood 就這樣做。大家今天談 Robinhood Chain 在 Arbitrum 平台上的成功，但一年前，不是 2026 年 7 月，而是 2025 年 6 月，Robinhood 先在 Arbitrum One 上線。過去一年他們在 One 上推出超過 2,000 檔代幣化股票，後來才搬到自己的鏈。Arbitrum 是唯一給你這種選擇的生態：先在 Arbitrum One 這種頂層生態鏈上線，到對的時間，用同一套可客製的技術，再去做自己的鏈。Offchain 會全程支援。以太坊生態最大的抱怨之一，是不知道該找誰。And by the way, that's well-precedented. Robinhood did this. Everyone talks about Robinhood Chain today and its success as a chain on the Arbitrum platform, but a year earlier, so not in July of 2026, but in June of 2025, Robinhood first launched on Arbitrum One. And they've launched over the course of the past year over 2,000 tokenized stocks on Arbitrum One, and then eventually moved to their own chain. And Arbitrum is the only ecosystem that gives you that optionality to launch on a top ecosystem chain like Arbitrum One, and at the right time, using the exact same technology that you can actually customize, go and build your own chain. Offchain, of course, is here to support you every step of the way. One of the biggest complaints we've had in the Ethereum ecosystem is that there just haven't been—people don't know where to go.

14:43別的生態往往只有一家公司。以太坊生態是分散的，人常常不知道要去哪裡建造。我今天要說：機構可以來找 Offchain，我們帶你走每一步，從鏈的基礎設施、營運、客製，到生態支援，再到錢包層的 ZeroDev。想帶走的重點是：2026 年若你在做區塊鏈協議或區塊鏈本身，二層在經濟和技術上都是最好的選擇。二層之中，Arbitrum 是做生意最好的地方。對參與生態的人來說，Arbitrum 本身也是一門收入夠分散的生意：鏈本身的費用直接進代幣持有人控制的 DAO、進鏈上錢包；Robinhood 這類鏈的費用也進同一個錢包；再加上昨天宣布的 USDG 孳息這條新收入。這些疊在一起，讓 Arbitrum 成為 2026 年做區塊鏈應用最好的生意之一。In other ecosystems, you know, there's only one company. The Ethereum ecosystem is distributed. People often don't know, "Hey, where do I go to actually build in the Ethereum ecosystem?" And I'm telling you today, you can come to Offchain as an institution, and we can walk you through every step of the process, from the chain infrastructure to running the chain infrastructure, to customizing the chain infrastructure, to ecosystem support, to wallet-level support with our ZeroDev platform. The takeaway that I'd like to leave you with today is, if you're building a blockchain protocol or a blockchain itself in 2026, L2s are the absolute best choice for you, both economically and technologically. And among L2s, Arbitrum is the absolute best place to build your business. And Arbitrum itself, for those participating in the ecosystem, is itself a thriving business with incredibly diversified revenue streams, from the chain fees itself, all going directly to the Arbitrum tokenholder-controlled DAO, the Arbitrum tokenholder wallet on-chain, to the fees from chains like Robinhood going to the Arbitrum tokenholder wallet as well, to new revenue lines like the USDG yield announced yesterday, all compounding to make Arbitrum one of the best businesses in 2026 to build your blockchain application on.

15:52非常感謝，謝謝各位的時間。Thank you very much. I appreciate your time today.