Written by: Jeff Yan

Compiled by: Wu Shuo Blockchain

This episode features a transcript of an interview with Jeff Yan, founder and CEO of Hyperliquid, on the When Shift Happens podcast. Jeff reviews the wealth effect and shift in responsibility brought about by TGE (Token Generation Event), explains Hyperliquid's core design philosophy of "no internal exchange, no discretionary power," and why it insists on automatically buying back and burning protocol fees rather than manually timing the process.

He emphasized that Hyperliquid is not a "crypto company," but rather a "financial protocol" that upgrades financial infrastructure with cryptographic technology, aiming to "housing all of finance"—enabling all financial activities to be completed within a composable, permissionless, and transparent on-chain system. Jeff also provided a detailed analysis of how key modules such as HyperEVM, HIP-3 (Permissionless Perps), Outcome Markets, the consortium stablecoin USDH, Kinettiq staking, and Hyperlend lending collectively construct a decentralized, scalable, and composable on-chain financial system.

In response to external criticism, he emphasized that Hyperliquid is not a corporate entity and there is no "timed buyback" or "human intervention." All mechanisms are executed via on-chain logic. He also pointed out that the real competition lies not in short-term data, but in whether a trustless, globally accessible financial system can be established.

The entire dialogue presented a core proposition: in an era of accelerated AI development, if the financial system does not upgrade to an on-chain, programmable, and open architecture, there will be no place for humans in the future financial world.

On the wealth effect and long-termism

Discussing the entrepreneurial environment in Singapore

Kevin: What do you think of Singapore?

Jeff: It's really great. It's definitely a great place to get things done and work on projects. It's very safe, modern, and the infrastructure works smoothly. Many people say it's "boring" and consider that a downside, but for me, it's actually a huge advantage. Here you can focus on building without too many distractions. It's a perfect place to bury yourself in your work.

Did they anticipate such a scale of wealth creation?

Kevin: Did you expect it to generate such a large-scale wealth effect at the time?

Jeff: I didn't expect it. To be honest, I rarely plan what the outcome will be at a specific point in time. We focus more on doing things to the best of our ability. We know our direction and what we're building, but I don't write things down, like "it would be great if we reached a certain quantifiable metric in a month." That's not my way of doing things.

As contributors, there are many things beyond our control. We need to focus on doing our jobs well. The outcome certainly depends to a large extent on our efforts, but it is also influenced by many external factors.

However, I was genuinely surprised by the overwhelming positive feedback that followed. It felt wonderful to hear people say, "This is rare in the crypto industry," and "This is what should be happening; we just rarely see real examples." Others said that Hyperliquid demonstrated a positive path that the industry could potentially take.

This gives me a great sense of satisfaction because I truly believe in this model and in the free market. There are many examples from the outside world that have been used to prove that "this model doesn't work." And when you become a case that can refute those doubts, it feels really good.

The crypto industry suffers from a talent mismatch and should "do the right thing" to change its image.

Kevin: Obviously, what you've done is incredibly difficult. But do you think there are enough people in the crypto industry trying to do the right thing in the right way?

Jeff: I think there's a general mismatch between what can actually be built and the people trying to build it in the crypto industry.

This largely stems from the crypto industry's long-standing negative reputation. You can compare it to fields like AI, or some traditional finance and technology projects before AI. If someone has great potential, graduated from a good school, or is exceptionally talented and wants to create something truly valuable—today, I think many of those people wouldn't even seriously consider entering the crypto industry.

Because this industry is full of negative cases, it has created a stereotype that it is only for people who don't take product development seriously and just want to make a quick buck.

But I think this view is completely wrong. This is not the real potential of this industry at all.

Kevin: So how can we change this situation?

Jeff: All we can do is keep building. Focus on building what we truly believe in.

When I say "we," I'm referring to the entire ecosystem, not just a single team. For example, the Hyperliquid ecosystem has never been too concerned with the so-called "market meta," nor has it been overly influenced by external opinions—such as what people theoretically believe "should be done now," or "which sectors are hotter." We are more like a community doing what we truly believe in.

I think this is probably the best thing we can do—use tangible results as feedback and demonstration.

Of course, as a permissionless network, it's inevitable that some bad examples will emerge. There will always be those who try to exploit the community and the mechanisms for short-term arbitrage or value extraction for personal gain. It's truly regrettable to see this happen.

Overall, I believe our community is on the right track. There's a great culture here. If we continue building in this way, hopefully the outside world will gradually notice, and more people who genuinely want to build long-term will be willing to enter the DeFi space.

The responsibility behind enormous wealth

Kevin: I think there are definitely a lot of people watching you right now. As a founder, when you create billions of dollars in wealth almost overnight, what do you feel comes with responsibility?

Jeff: I'm not sure if specific events like the Genesis or TGE really changed the nature of responsibility itself. I believe that building any financial product inherently involves enormous responsibility.

If you choose to work in financial infrastructure, you must commit fully. Even so, you'll still feel there's so much to do. This is because finance is an extremely important part of a person's life. The level of trust someone places in you when using financial products is far greater than when using ordinary consumer products.

So I don't have a particularly grand answer. All we can do is try our best to execute it, and we must do it almost perfectly.

This could mean slowing down; it could also mean choosing to build systems in a more difficult and complex way. There are many similar examples. We must always design systems around some core principles: neutrality, fairness, and robustness across various market environments.

These principles will not change because of TGE. Even before that, users already trusted Hyperliquid—trusted it as an on-chain, fully verifiable, and transparent alternative to the opaque systems they had grown tired of. It is this trust that is the real responsibility we need to shoulder.

From "insider faith" to global users

Kevin: Our last recording was about a month before TGE. At that time, I did something I often use to test the authenticity of the community—because there's a lot of fake data and bot accounts in the crypto industry, which is a well-known problem. So I took a group photo and posted it to see the community's reaction. The enthusiasm far exceeded my expectations, and I felt that this community was even stronger than I anticipated.

It felt a bit like an "insider's faith," but more like a hidden code—those who understand it know it, but many in the crypto community don't fully grasp what's happening. Then TGE happened, and almost everything was done correctly. You transformed from a "faith-based community" with strong cohesion within the crypto world into a project supported by the entire industry.

So, how do we make Hyperliquid a "code" that can attract users outside the crypto community? How do we get people outside the crypto industry to join in?

Jeff: I think, ultimately, it comes down to providing real value. This applies not only to Hyperliquid, but to all Builders.

I don't believe the goal is to create a "fervent community." What people perceive as a "sense of faith" or a "cult vibe" is essentially the resonance people feel with a group of builders who are moving towards a common goal and share clear values. It seems rare because such long-termism and alignment of values ​​are uncommon in the crypto industry.

I'm reluctant to call it a "faith." To me, it's simply people coming together to do the right thing in the right way, upholding principles of fairness, and building an open financial system. This should be the "normal" way.

If we want the whole world to join in, we must continuously demonstrate that we can build value that traditional financial systems cannot. I am very confident about this.

From a broader perspective, DeFi has always embodied this promise. Past issues have primarily stemmed from the gap between the concept and its technological implementation. The benefits of an open financial system to the world are virtually undisputed. The key lies in implementation and preventing unethical behavior, short-term temptations, or personal gain from undermining the overall direction.

I believe there are only two core points: first, effective execution; second, consistently doing things the right way. If these two points can be achieved, then DeFi is essentially a massive technological upgrade to the existing financial system. When this becomes clear enough, people will naturally join in.

Will a milestone be celebrated?

Kevin: Did you celebrate at the time? Or was everything completely normal? I'm asking this because I remember a story, I think it was Matt Huang from Paradigm who mentioned it—a Web2 company went public, and they prepared champagne for the day, but the whole team just drank a glass of water and went back to work until midnight, with almost no celebration.

So I'm curious, what happened at Hyperliquid Labs back then? You just said it was just a milestone, and there's still a lot to do. Did everyone have some kind of celebration? Or was it, "This was all part of the plan," and then they moved on to the next goal?

Jeff: I've been thinking about it... and it seems like we didn't really have a formal celebration. Looking back now, it's a little regrettable. Because if you were to reflect, there were definitely many moments worth celebrating. But in the moment, there's always more to do. You can never find the "perfect moment" to stop and celebrate.

Perhaps it will be many years from now, when the entire system operates completely autonomously, everything is mature and complete, and it truly becomes a stable financial system, that will be the time to celebrate.

Of course, everyone may celebrate in their own way. But at least at Hyperliquid Labs as a team, we don't have that "pop champagne every time a feature launches" culture. That's not our style.

I think this is more of a cultural issue—we are naturally more focused on what to build next, rather than stopping to celebrate what has already been accomplished.

Regarding team selection, workload, and privacy boundaries

How to screen "highly trustworthy" members

Kevin: How do you test whether someone has "high integrity"?

Jeff: That's a very good question. Frankly, there's no perfect way.

In Hyperliquid Labs' hiring process, we certainly conduct extensive technical assessments, but beyond that, we always allocate at least a full day for real-world collaboration. Not the kind of highly time-pressed, simulated interviews, but genuine working together.

When you collaborate with someone for an extended period, discuss problems, write code together, and face real-world challenges together, you gradually get a feel for what kind of person they are. There are many "soft signals" involved—difficult to quantify, but perceptible. You can never be 100% certain about someone's character, but sometimes signals will appear that make you feel, "This risk isn't worth taking."

Additionally, we ensure that everyone on the team feels genuinely comfortable with a new member. There have been instances where some people are very supportive of a candidate, while others have reservations. In these situations, we typically choose not to hire them.

Kevin: So it's a consensus mechanism? Everyone votes together?

Jeff: It wasn't exactly a formal vote. There wasn't much of a structured voting process. But basically, if even one person voiced a moderate to strong objection, that was usually enough to veto the hiring decision.

Regarding work intensity and sleep

Kevin: How much sleep do you get? How much sleep do your team members get? How much sleep do you "allow" them to get?

Jeff: It really varies from person to person. We don't put pressure on "time commitment" because that's usually the start of a disaster.

If someone truly needs sleep but doesn't sleep well, the quality of anything they produce will decline. Our quality standards are extremely high—even when fully energetic, writing excellent engineering code requires our utmost effort.

Therefore, we operate on a highly trust-based system in this regard. I also don't really agree with the "all-night culture" or the atmosphere that glorifies staying up late. I personally work longer hours than many who advocate this culture, but I don't think that should be a rigid requirement for the team.

It's true that some people on the team work very late, and others have fixed, highly productive periods each day. As long as they consistently produce A+ level results, we don't care how many hours they actually spend.

For us, what matters is not "how long we work", but "the quality of our output".

Employee Token Unlocking and Privacy Boundaries

Kevin: One last question about the team. I promise this is the last one. There's been a lot of FUD (Fear, Uncertainty, and Doubt) over the past few months, and one of the concerns from community members and token holders is the employee token unlocking schedule. Many people are worried that approximately $300 million worth of tokens will continue to flow into the market every month for a long time. What's the actual situation?

Jeff: We won't discuss these specifics publicly. Because I genuinely believe that financial privacy is a fundamental right. That's one of the reasons we entered the crypto industry and got involved in DeFi.

Just as you wouldn't ask a community member to publicly state, "You hold a lot of tokens, and you must tell everyone what you're doing with every single transaction," I think there needs to be a clear line drawn here, and that line is personal privacy.

I believe that how the protocol operates must be completely transparent. The flow of every dollar should be clearly traceable; assets within the system must truly belong to their respective holders; and the protocol rules must be public and verifiable. This level of transparency is an uncompromising prerequisite for an open financial system. Without this transparency, the problems we see in some centralized exchanges will arise.

On the other hand, how any contributor handles their own tokens is their personal asset. This should not be subject to public scrutiny, nor is it my responsibility to judge or disclose it.

Agreements must be transparent, but individuals must have privacy. These two are not mutually exclusive.

Regarding addressing FUD, automatic destruction mechanisms, and eliminating human decision-making power.

How to deal with FUD and external attacks

Kevin: Your approach to handling FUD and external criticism is quite interesting. Many people think that Hyperliquid's approach is to ignore it and remain silent.

Jeff: Actually, that's not the case. It's true that in the early stages we tended to not respond. But later I realized that my own intuition about PR isn't always correct. So now I prefer to leave that part to more professional people to judge and handle.

One thing we later consciously tried to adjust was that when FUD (Fear, Uncertainty, and Uncertainty) occurred, my instinctive reaction was often: "This is obviously wrong, and the truth will eventually come to light." But I gradually realized that this mindset is not always correct.

From a pragmatic perspective, if there is false information circulating and clarification is indeed necessary, we will respond. We will address and correct specific issues, rather than ignoring them entirely.

Of course, we won't respond to every voice. The key is to determine: will this mislead users, damage community trust, or affect the proper understanding of the agreement? If so, we will address it directly.

Therefore, rather than saying we "ignore FUD," it's more accurate to say we are learning how to deal with it in a more mature and strategic way.

Kevin: Will these FUDs have any impact on you personally? To be honest.

Jeff: It depends on what type of FUD it is.

The crypto market has experienced significant volatility over the past six months, with much of the time trending downwards. Certain events have indeed caused substantial damage to the market. And because Hyperliquid is one of the few truly transparent trading venues, much of the discussion has naturally focused on us.

I was genuinely concerned during that period. I was almost acting with a sense of "mission," trying to explain some technical details clearly. Because these issues are very personal—many people may have actually lost a lot of money. In that situation, emotions are highly sensitive.

However, I've also noticed that some competitors are deliberately "misleadingly portraying" Hyperliquid from different angles. They downplay what we did right and use negative narratives to divert attention from their own problems.

This situation makes me angry. Because it's obvious they know what they're doing.

On the other hand, this anger eventually turned into motivation—driving me to explain more clearly to users why Hyperliquid is designed this way and why transparency is so important.

For example, some centralized exchanges cite statistics from certain data analytics websites, saying, "Look, Hyperliquid has more liquidations than us." But if you look at those platforms' documentation (though few people actually do), you'll find that they may only report "the first liquidation per second," without disclosing the full data. There's no explanation for why only this part is included in the statistics.

On Hyperliquid, every order, every cancellation, and every liquidation is publicly verifiable on the blockchain. Comparing partially disclosed data with fully disclosed data will inevitably lead to misleading conclusions.

This is essentially information asymmetry, or even blatant misleading.

What's even more frustrating is that when influential voices with large followings in the industry amplify this narrative, the impact is amplified rapidly.

Our voices may be quieter and our reach narrower. But precisely because of this, we need to explain the facts loudly, clearly, and repeatedly at crucial moments.

I won't pretend these things don't affect me. They do. But they bring more responsibility than retreat.

Automatic buyback and burn mechanism vs. discretionary buyback

Kevin: There are many people in the ecosystem who support you, but there are also some quite intelligent people in the industry who are criticizing Hyperliquid. A common argument is that you are buying back tokens every day, but you should stop buying back when the price is high and buy back more when the price is low. What do you think is the core problem with this argument?

Jeff: The core issue is that we are not a discretionary buyback program at all.

Many people interpret this as a company action, but this is a flawed analogy. Hyperliquid is a rules-based protocol, not a company that can make subjective decisions.

To draw an analogy: On Ethereum, priority fees are now directly burned. You wouldn't ask Ethereum developers to say, "When the price of ETH is high, Vitalik should use these fees for other investments instead of burning them."

Because that's not a decision that can be made subjectively—it's part of a protocol-level rule.

The same applies to Hyperliquid. Fees generated by the protocol are automatically converted into HYPE according to preset rules and then destroyed. This is not a matter of the team deciding "how much to buy today and when to stop" on a whim. The entire conversion logic is written into the chain's execution mechanism.

I believe this misunderstanding arises because many centralized or semi-centralized exchanges do indeed have token buyback programs. In that model, the company can decide when to buy back, how much to buy back, and whether to suspend the process. For that structure, discussing timing-based buybacks is reasonable.

However, on Hyperliquid, the conversion of fees into HYPE and their destruction is an automated on-chain logic.

This is similar to on-chain TWAP order execution—when you submit a TWAP order, no one sits there manually deciding when to split the order or when to send the next order to the order book. That's part of the protocol execution logic.

Similarly, how fees are converted and destroyed is entirely part of the on-chain execution logic, not human intervention.

Therefore, the crux of the problem is that many people misunderstand a "protocol-level automatic destruction mechanism" as a "company-level timing buyback strategy." These two are completely different in terms of design philosophy and governance logic.

Why must human decision-making power be eliminated?

Kevin: Why is "no discretionary power" so important to Hyperliquid?

Jeff: Because Hyperliquid is not a company. This is another common misconception.

When I say "we," or refer to Hyperliquid Labs, I mean we are a very small team. Our role is to build a small but crucial component of the entire Hyperliquid ecosystem. We are not "Hyperliquid itself."

Many people are accustomed to using analogies from traditional finance to understand everything. This is because people don't like learning entirely new concepts; they prefer to fit new things into existing frameworks. While this may lead to some similarities, it also creates many areas where it's simply not applicable.

Hyperliquid is not a mapping of any existing entity in the traditional financial world. It is not a company, not an exchange, and not an organization that "creates value for shareholders."

Our vision is to enable the financial system itself to operate on the blockchain.

Yes, the protocol does create value and returns that value to the native token. That's positive. But it's a completely different logic from a company rewarding its shareholders with operating profits.

A company is a centralized entity with management, a board of directors, and subjective decision-making power. Hyperliquid, on the other hand, is a neutral platform, an infrastructure upon which finance is built.

From another perspective, it's more like the internet's relationship to information. The internet isn't a company, nor is it designed to generate profits for "internet shareholders." It's a set of neutral infrastructure protocols that allow information to flow freely.

Similarly, Hyperliquid, as an ecosystem, network, and protocol, aims to become the infrastructure layer in the financial sector.

Once you grant a core team "discretionary power," such as deciding when to repurchase or when to suspend, you are essentially bringing the system back to a "company model."

We chose a rule-based, automated, and unchangeable mechanism precisely to ensure that the system is neutral and trustworthy, rather than relying on the judgment of a particular team.

If the goal is to build a truly on-chain financial system, then eliminating human decision-making power is not an option, but a prerequisite.

Regarding HyperEVM, Permissionless Perps, and On-Chain Finance

The essence and significance of HyperEVM

Kevin: Last time you were on the show, we talked about your ambition to become the "AWS of liquidity." This time, let's discuss in detail what has happened recently. Let's start with HyperEVM. What is HyperEVM? Explain it in the simplest way possible, without using industry jargon.

Jeff: In its simplest terms, HyperEVM allows developers to migrate their smart contracts, which were originally deployed on Ethereum, directly to Hyperliquid and run them with minimal modifications.

You can think of it as a "blank canvas." Developers can deploy contracts and run logic on it, just like they do on Ethereum.

On the surface, it looks very similar to Ethereum in terms of user experience. The innovations that Ethereum has accumulated over the past few years—development tools, contract standards, and application models—can all be reused on Hyperliquid.

Kevin: What are your honest thoughts on the current performance of HyperEVM?

Jeff: That's a somewhat delicate question.

First, I think one misconception about HyperEVM is that it "didn't succeed." I could give many examples of how well it has done if needed.

But more importantly, if you look at it from another level, HyperEVM is not just "borrowing the Ethereum framework." Its true uniqueness lies in the fact that it allows smart contracts to directly call Hyperliquid's native on-chain capabilities for the first time. This is crucial.

I feel that this narrative hasn't been fully understood yet. HyperEVM isn't just another standalone EVM chain. It's more like an "entry point"—an interface that allows direct connection to Hyperliquid's native liquidity and on-chain infrastructure.

Therefore, when evaluating HyperEVM, it shouldn't be simply compared to other completely self-contained EVM chains. It's not an isolated ecosystem, but rather a gateway to Hyperliquid's core capabilities. From this perspective, its significance is entirely different.

Why is HyperEVM undervalued?

Kevin: Simply put, is it that people haven't truly understood HyperEVM yet?

Jeff: I think some people understand it, but this understanding hasn't truly entered the mainstream narrative yet. For example, a long-standing criticism has been: "Why isn't Circle deeply integrated yet?"

The problem is that HyperCore itself is a completely native implementation of a whole set of financial primitives, such as account balance ledgers. This architecture is quite different from the structure of traditional general-purpose public blockchains. For companies that already have mature infrastructure specifically adapted to general-purpose public blockchains, directly connecting to HyperCore is actually very difficult. This is where the significance of HyperEVM lies—it provides a bridge.

For example, the minting and burning logic of native USDC can be accomplished through Circle's existing Hub-and-Spoke network structure: you can burn on one chain and mint on another. HyperEVM is one of the supported chains.

But to truly understand this, you must understand the purpose of the entire architecture design. It's not simply about "deploying an EVM environment," but about enabling external infrastructure to connect in a standard way while also allowing for deep interaction with HyperCore's native capabilities.

The problem is that when these integrations are done well enough, they are "unnoticeable." The user experience is so smooth that you don't realize it's achieved through the special interaction between HyperCore and HyperEVM. Because of this, it's difficult to create a narrative of an "explicit success story."

Many people ask, "Why haven't I heard any success stories about HyperEVM?" On one hand, many truly large applications are still being built. On the other hand, those successful integrations, precisely because they are so native and natural, don't seem "special." If an infrastructure is truly successful, it seems like it should be there for all the time. This, to some extent, is also a reason why it is underestimated.

HIP-3: The Significance and Challenges of Permissionless Perps

Kevin: Another thing you launched last year might not have been so easy for most people to understand at the time, but now we're seeing very concrete results with HYPE's HIP-3 — Permissionless Perps —. What exactly does that mean? Could you explain it in the simplest possible way possible?

Jeff: Simply put, perpetual contracts are one of the most important innovations in the crypto industry. While the theoretical basis for perpetual contracts may have existed in academic literature for some time, it was BitMEX that truly succeeded in putting them into practice and proving them as a more efficient price discovery tool. It concentrated liquidity on a contract that never expires, thereby improving market efficiency.

The core concept of HIP-3 is actually quite simple: there are no "unique" limitations in the crypto industry when it comes to perpetual contracts. Therefore, as long as an asset is liquid and already has a futures or options market, there should also be a corresponding perpetual contract. We are not saying that perpetual contracts are necessarily better than all other derivatives, but users should have this choice.

I've always believed that if traditional finance adopts perpetual contracts more widely across more asset classes, a significant portion of market trading volume will migrate to perpetual contracts. This is because they are among the most direct and efficient tools for expressing opinions, especially in terms of leverage and two-way trading.

In short, HIP-3 allows anyone to deploy perpetual contracts on Hyperliquid, also known as "Permissionless Perps".

But behind this statement lies a great deal of complex work. Because before this, no one had ever truly built a completely permissionless perpetual contract deployment system. For example, there are many core issues that need to be addressed:

• First, can the agreement delegate critical functions like Oracle real-time updates to a license-free deployment system?
• Second, perpetual contracts are highly coupled with the margin system. Once they become permissionless, a large number of edge cases and risk scenarios will arise, requiring very rigorous design.
• Third, how can we maintain system robustness and security while ensuring open deployment? These issues cannot be solved simply by changing a few lines of code; they are deeply intertwined. Furthermore, it requires builders to truly invest time in understanding a completely new mechanism and be willing to build applications on top of it. This itself also takes time.

Frankly, before HIP-3 was launched, many people probably thought it might not succeed. We ourselves couldn't be 100% certain that it would succeed. But logically, it should be feasible.

Moreover, when an ecosystem gathers a large number of capable developers and users who truly identify with the ecosystem and are willing to help launch new network effects, then such innovation should succeed in such a place—an environment like Hyperliquid, which has the fertile ground for success.

The fundamental difference between Hyperliquid and centralized perpetual platform models

Kevin: Can you explain what you're doing and how it differs from what everyone else is doing?

Jeff: I'm not in a good position to comment on what "everyone" is doing, but many of the attempts I've seen to build perpetual contracts (Perps) in non-crypto contexts follow this basic logic: an exchange that was already running a crypto perpetual contract business says, "Let's connect to price data sources (feeds) for other assets and then provide perpetual contracts for those assets in the same way."

This model can be seen in many centralized exchanges, as well as some decentralized exchanges. The core idea is to use the same platform, the same architecture, and the same team to list perpetual contracts for a wider range of assets.

Kevin: So what's the real difference between your approach and those of these players? Why should users care?

Jeff: For end users, it's not necessarily necessary to care about how it's implemented at the underlying level. But if we really want to move the entire financial system onto the blockchain, then we must have "domain experts" who truly understand a particular field to deploy and build these products.

If you look at the traditional financial system, each country has its own exchange system; under the country, there are different asset classes; different asset classes are traded on different exchanges with different rules; even within the same country, different asset classes have different trading interfaces. Continuing to break it down, you'll find that the financial industry is inherently highly specialized.

Of course, some of this complexity does stem from the inefficiencies of the traditional financial system itself. The crypto industry has, to some extent, eliminated some of the intermediaries. But on the other hand, finance itself is too large to be entirely built by a single team.

Many large financial companies may choose an alternative path: continuing to maintain a centralized "super platform," attempting to integrate users' entire financial lives into one place. They might do this through acquisitions, partnerships, or internal development, integrating all asset classes into a unified system. This is a top-down integration approach.

This model does indeed advance faster, is less fragmented, and may offer a better initial user experience. However, the problem is that it is very fragile. Once problems arise, it is often because power is excessively concentrated in the hands of a single entity. When a system centralizes everything, the risks are also centralized.

I believe that end users will choose the system that truly empowers them with financial sovereignty. This system must be globally accessible, have sufficient liquidity, and possess all the functionalities they require.

To build a system that truly serves global users, decentralization must be achieved across multiple dimensions. Decentralized ownership is one important principle, and it's also one of the core principles emphasized by Hyperliquid. But equally important is the decentralization of who builds the core products that users actually interact with.

This is the fundamental difference between us and other trading platforms: instead of a centralized team listing perpetual contracts for all assets, we build an open system that allows builders from different fields to deploy markets they truly understand and excel at on the same underlying protocol.

Why must all finance be brought onto the blockchain?

Kevin: Please explain in one sentence why it is so important to move the entire financial system onto the blockchain.

Jeff: Because if we don't do this, the financial system will fall behind the pace of technological development. The world is changing too fast, and the traditional financial system cannot keep up with this pace or meet the needs of users in a rapidly evolving environment.

Kevin: Why is Hyperliquid's approach better?

Jeff: This isn't just Hyperliquid's approach. Hyperliquid is simply putting into practice the values ​​that DeFi has always upheld. We are just a community trying to implement these ideas in a way that has a greater chance of truly succeeding and achieving mass adoption.

These core values ​​include: the financial system should not be controlled by companies, but should be a neutral foundational layer, like the internet.

You just mentioned AWS, which is an interesting example. Although AWS is controlled by a company, in a sense it's like neutral infrastructure. Nobody says, "Why did you choose AWS? It's a choice with a bias." It's more like the default infrastructure.

The financial system should also possess this attribute—to become a neutral, globally accessible infrastructure. People should not be forced to undergo layers of approval and overcome numerous barriers to access the financial system because of their nationality, identity, or whether they come from a developed country.

Furthermore, the financial system should be deeply programmable. This doesn't mean a fintech model that adds "programmable functions" to a traditional, non-programmable system, but rather one built from the ground up on programmable primitives. In other words, the basic components themselves are code, smart contracts, and can be directly interacted with by humans, intelligent agents, and various programs.

If finance is to keep pace with technological advancements, it must become a neutral, programmable, and globally open infrastructure, much like the internet. This is the fundamental reason why we believe all financial transactions must be brought onto the blockchain.

HIP-3 Data Performance and Silver Trading Cases

Kevin: If this approach is better, then the data performance should be better as well. Can you provide us with some key data to prove that Hyperliquid is operating the perpetual contract market the right way?

Jeff: I want to emphasize again that we are still in a very early stage. The future is never certain, and short-term data alone cannot prove long-term success.

That said, existing data already demonstrates the power of permissionless protocols to some extent—the effects can be quantified when you allow deployers the freedom to express their capabilities creatively and truly execute their vision.

A prime example is the silver market. Recently, some metal assets, including silver, have experienced extremely volatile fluctuations, with several consecutive days of price action resembling "10 Sigma." Both institutional and retail traders are paying close attention to this market.

On Hyperliquid, the related marketplace launched via HIP-3 (primarily operated by the first deployer XYZ) already accounts for approximately 2% of global trading volume in silver price discovery.

2% may not sound like much, but considering the size of the global silver market, it's a truly astonishing figure. More importantly, these HIP-3 markets have only been online for a few months. For a newly launched on-chain perpetual market, this is incredibly impressive growth.

Moreover, this achievement wasn't due to the core team "doing the trades." The core team's job was to build the underlying primitives—HIP-3 itself does require a lot of engineering work. But how these liquidity infrastructures are ultimately used depends entirely on the developers who choose to build on top of them.

This is more like proof of the builders' quality: they chose Hyperliquid because only this platform allows them to implement the market structure they want to build.

Kevin: You just said that the data is starting to get interesting, but it's still very early, so the future may not necessarily follow this direction. Why do you say that?

Jeff: Because it's very dangerous to assume an idea is "unshakeable" just because it currently seems to have supporting evidence. The world is changing too fast.

Some principles are relatively stable, such as: the financial system should be open to everyone globally; finance, like information, should exist within an open system; and users should have financial sovereignty. These principles are not easily changed. However, any specific implementation method, if overly rigidified, becomes fragile.

When designing the core primitives of the blockchain, an important starting point is that when builders and users from different directions are all reporting the same problem—"We want to do something, but the current system can't"—if you find that these feedbacks have an intersection, then that intersection itself is often a basic capability worth writing into the protocol layer.

When the protocol layer makes very restrained and selective decisions about "what must become native primitives and what should be left to the general EVM or application layer," the system becomes more elegant and resilient.

In this way, the protocol itself does not overly embed any particular business model or market judgment, but rather leaves the "decision-making with a stance" to the builders. The builders can quickly adjust according to real market changes, while the underlying layer remains neutral.

All HIP proposals, including HIP-3, are based on this kind of thinking.

Spot trading becomes a price discovery center

Kevin: We discussed the Spot Trading feature before. It wasn't developed by the core team, but rather by a team called Unit, which plans to build it in 2025. What does this tell us?

Jeff: In 2025, there were several very crucial project launches, and in the early stages of these assets, Hyperliquid became the primary source of spot price discovery. Moreover, this is the first time to my knowledge that a decentralized exchange has become the dominant spot price discovery center for a new asset.

One example is XPL. At the time, Hyperliquid was leading in trading depth and volume in the on-chain spot market during a key phase.

In a sense, spot trading is the form closest to Satoshi Nakamoto's original vision of "peer-to-peer electronic cash." This is because when you engage in spot trading, you are essentially transferring assets on the blockchain ledger.

One party transfers its quote asset to another; the other party then transfers its base asset back to the first party. Order books, local matching mechanisms, and similar mechanisms are merely coordination tools that facilitate these on-chain asset transfers.

If you asked Satoshi Nakamoto how Bitcoin should be traded, I'm very confident he would say it should be done entirely on-chain. Of course, the Bitcoin network itself may not be well-suited to handling complex matching logic, but the transactions themselves should be peer-to-peer.

We've done a lot on Hyperliquid, but this concept of "native on-chain execution" is actually a very core part of our DNA.

I believe spot trading is a particularly good testing ground. When all the pieces are in place—performance, order book, user experience, liquidity—users are actually more willing to trade in a way that aligns with the technical logic of crypto assets themselves.

Most people trade spot crypto assets because they believe in the vision of crypto and DeFi. After many years, finally being able to trade their own assets entirely on-chain without sacrificing user experience feels like completing a closed loop.

From technological ideals to practical implementation, and then back to the original point-to-point concept—it feels like "coming full circle and returning to the starting point." I think that's really cool.

On early-stage startups, non-linear financial primitives, and on-chain financial systems

Jeff's early experience with failed prediction market projects

Kevin: You posted some videos on X before, saying that "young Jeff" did a prediction market project in 2018, around the same time as Kalshi. Why didn't it succeed back then?

Jeff: There are many reasons. The simplest and most honest answer is—we weren't ready at the time.

The idea itself was actually quite cool. From an infrastructure perspective, we were on the right track in many ways, such as using off-chain matching and on-chain settlement. Given the technological limitations at the time, it was a fairly reasonable architectural choice.

But truly making a product successful is far more complex than simply "having the right idea." In fact, the idea itself accounts for a very small percentage of the entire process. Execution, product refinement, market timing, user education, liquidity activation, and psychological resilience are all equally important.

At the time, it wasn't obvious to me that anyone actually wanted to use on-chain products. When the project first started, market sentiment was high, prices were rising, and everyone was excited. But in the second half of 2018, the market took a sharp turn for the worse, and prices fell sharply.

We have repeatedly witnessed this cycle in the crypto industry—users flock to it during bull markets and almost no one cares during bear markets.

But if that's your first time experiencing this cycle, and you go to communicate with users only to find that almost no one is willing to try your product, the blow can be devastating, especially when you've invested a lot of time and energy.

Even so, that experience was very educational. We truly understood what kind of long-term investment and mindset are needed to build a successful financial product.

In a sense, the idea didn't disappear. Later, some teams that started around the same time as us persevered, such as Kalshi. They clearly had the conditions and resilience needed to make things happen, and now they are beginning to see the results of their years of hard work, as well as wider mainstream adoption.

Seeing these companies gradually achieve results makes me genuinely happy for them. This in itself shows that many ideas are not necessarily wrong; it's just a matter of whether you are prepared to persevere with them at the right time and in the right way.

Outcome Markets: Non-linear Expression Tools

Kevin: Eight years later, you're back to prediction markets. But this time, a more accurate term would be Outcome Markets. What's the difference?

Jeff: This goes back to the "primitives" design philosophy we discussed earlier. On Hyperliquid, we wanted the protocol layer primitives to be as small and self-consistent as possible—simple in structure, clear in performance, but with the widest possible range of applications.

Outcome Markets' premise is that it will become a core tool for people to express "non-linear perspectives" on the blockchain.

Let's compare: Spot is the most basic form of asset. You hold the asset, and the transfer is completed on the blockchain ledger when trading. This is a fundamental primitive that any financial system should have, and there is no dispute about it; Perpetual contracts are relatively more controversial, but more and more people recognize that their value in price discovery and capital efficiency is significant enough to warrant being part of a general financial instrument. HIP-3 can be seen as a phased realization of this idea.

However, despite the significant differences in their mechanisms—perpetual trading allows for leverage and higher capital efficiency, while spot trading represents the tokenization of real assets.

They are similar on one key point: they can only express a "linear view." Whether you use 1x leverage or 100x leverage, essentially your profit or loss changes linearly with every unit price movement. For many users, this is sufficient. But some users actually want "non-linear results."

Here are a few examples: You might want a contract to result in either $1 or $0, depending on whether a certain event occurs (a typical binary structure); you hold Bitcoin and are generally bullish in the long term, but if the price falls below a certain level, you want downside protection; you might want a structure where the profit is limited to a certain range. These are all non-linear perspectives. They cannot be precisely expressed by a combination of spot and perpetual contracts alone. This is the significance of Outcome Markets.

Outcome Markets are essentially "fully collateralized contracts." Both parties deposit collateral in advance, and the final settlement is redistributed between them based on the outcome. The settlement result can be binary or continuous, and crucially, it eliminates liquidation risk. Users deposit capital and await settlement; there are no complex mechanisms such as forced liquidation, margin maintenance rates, or funding fees.

Kevin: For example, options and market prediction are similar examples?

Jeff: Yes, options and prediction markets are the most intuitive examples. They are both non-linear payoff structures.

In addition, there are some more interesting applications. For example: Can the market aggregate opinions instead of just trading prices? Can important issues be brought to the surface through market mechanisms? Can a capital-driven consensus be formed around a certain subjective outcome? These can all be naturally categorized into the framework of Outcome Markets.

Perpetual contracts will remain an important price discovery tool, but they belong to a relatively complex financial primitive; while Outcome Markets offer a more basic and general non-linear expressive capability.

In simple terms: spot trading solves the problem of "owning assets," perpetual trading solves the problem of "linear viewpoints + leveraged expression," and outcome markets solve the problem of "non-linear viewpoint expression." Together, these three constitute a more complete on-chain financial expression system.

The true meaning of "Housing All of Finance"

Kevin: You often mention "housing all of finance." What does that actually mean?

Jeff: Its core meaning is that the financial system shouldn't be fragmented. That's what "all" means. You shouldn't need to open a bank account first, and then transfer funds to another account if you want to buy something, or open a new account to participate in a product. Ideally, a user's entire financial life should be completed in one place. I think "composability," that is, integrating users' financial activities into a unified system, is a huge breakthrough.

However, this isn't unique to the crypto industry. The crypto industry's concept of "housing all of finance" is more importantly about the platform itself not being controlled by a single centralized company. It should allow many different teams to participate in building. When someone builds a new application or protocol, it should automatically connect and interact with systems built by other teams. This network effect is what truly constitutes a financial system.

Throughout this process, there has always been a trade-off: on one hand, there is a fragmented but highly composable system; on the other hand, there is a system that may be more efficient but is operated by a centralized institution and has relatively limited functionality. There has always been tension between these two.

However, with proper design, it is possible to strike a balance between the two ends—maintaining composability while also considering efficiency. This concept actually drives a lot of our thinking in product and system design.

Hyperliquid is not a crypto company, but a financial protocol.

Kevin: You told me before that Hyperliquid isn't a cryptocurrency company, but a financial company that uses cryptography as its underlying infrastructure. Of course, the word "company" might not be entirely accurate. Could you explain that?

Jeff: I think it's a matter of distinction. We genuinely believe in DeFi values. From that perspective, the ideas of crypto and DeFi are ingrained in our DNA.

But we don't define ourselves as a team that "builds the best crypto exchange." That's not our vision.

In my view, finance is fundamentally finance. The significance of crypto lies in its potential to upgrade the underlying "rails" of people's financial activities. That is, to reconstruct financial infrastructure with blockchain technology, allowing people to complete financial transactions in a more open, efficient, and transparent system. This is the direction we are building.

If the Hyperliquid ecosystem ultimately succeeds, some might say, "This proves Crypto is right; Crypto has been validated." I think that's a reasonable interpretation.

But some might say, "This is a completely new thing that borrows some ideas from the crypto space, but has eventually developed into an independent system." I think this interpretation is also valid.

The key is not the label—whether it is a "crypto company"—but whether it truly drives the upgrading of financial infrastructure.

USDH Alliance Stablecoin Mechanism

Kevin: Hyperliquid currently has only 11 core team members, but it also boasts an ecosystem of some of the smartest developers and builders in the industry. Let's talk about one of your projects. I noticed you have a stablecoin called USDH. Why did you create this?

Jeff: About six months to a year ago, the entire industry experienced a surge of interest in stablecoins across multiple blockchains. At that time, the "Genius Act" was passed, and the market generally believed that stablecoins would become the entry point for institutional funds to enter DeFi rails for the first time, so everyone had a lot of interest in stablecoins.

Within the Hyperliquid ecosystem, we proposed a protocol-level concept called "Alliance Stablecoin." This mechanism wasn't predetermined from the beginning; rather, it was a design ultimately reached at the protocol layer after repeated discussions with many stakeholders in the ecosystem.

The core of this design lies in the fact that the returns behind the stablecoin (such as returns from assets like government bonds) can be distributed between the protocol itself and the stablecoin issuer according to preset rules. USDH is operated by the Native Markets team, whose members include experienced founders with institutional backgrounds and some very strong ecosystem community builders. I think they are an excellent team, and I look forward to them building a truly deeply integrated and aligned stablecoin on Hyperliquid.

These stablecoins have a highly symbiotic relationship with applications and developers within their ecosystem.

For example, the protocol itself generates revenue, thus HYPE token holders are aligned with the growth of this stablecoin; simultaneously, users who transact using this stablecoin enjoy lower transaction fees. All these rules are written into the protocol layer and are entirely objective, automatically executed mechanisms, not human decisions.

We believe this design creates a very strong synergy and has the potential to unlock a whole new set of application scenarios.

Kinetiq's full-chain staking innovation

Kevin: What is Kinettiq? Why is it important to the Hyperliquid ecosystem?

Jeff: Kinetiq is currently the largest liquidity staking protocol on Hyperliquid. It embodies a key part of the HyperEVM vision.

As we mentioned before, HyperEVM can directly "read" and "call" HyperCore's capabilities through a specific mechanism, essentially opening a window for smart contracts to the underlying core system, allowing them to use the core functionality as a tool. This architecture enables liquidity staking to be built entirely on-chain.

Typically, staking is built on a system mechanism at a lower level than the application layer. Taking Ethereum as an example, although mainstream liquidity staking tokens (such as stETH) are on-chain tokenized assets, they often have certain off-chain components or operational processes behind them. In other words, the tokenization process is not completely closed-loop and completed on-chain.

On Hyperliquid, staking-related read and write operations—including staking, unstaking, and yield calculation—can be directly invoked by smart contracts through precompiles and core writer primitives. This means that a smart contract can tokenize a pool of staked HYPE tokens, just as it does on the EVM; simultaneously, the entire process of staking, unstaking, and accounting is completed in a closed loop within the same on-chain system. This is a truly "fully on-chain" liquidity staking mechanism.

Its significance lies in eliminating the risks associated with bridging and additional trust assumptions. There are no off-chain dependencies or additional custodian links; everything is guaranteed by the protocol layer. Therefore, we believe this is a very cool and important innovation in the field of liquidity staking.

The key role of Hyperlend in the Hyperliquid portfolio margin system

Kevin: What Hyperlend is building, why is it important to Hyperliquid?

Jeff: Hyperlend is currently the leading lending protocol on the EVM. Since its launch, its TVL has seen significant growth, which I believe is mainly due to their continuous iteration and solid execution. It currently runs entirely on the EVM; users can deposit collateral as suppliers, and borrowers enter the market to borrow assets. Essentially, it's a market mechanism where interest rates are determined by supply and demand, with borrowers paying interest to suppliers.

This is a fairly common DeFi model, but this lending system will become very important once Hyperliquid's HyperCore Portfolio Margin feature moves beyond its early stages.

We haven't discussed portfolio margin in detail before. Its core concept is that traders want to be able to use any assets they hold, especially highly liquid assets (such as BTC), as uniform collateral and trade in any market based on these assets, rather than being restricted to a specific market. This reduces liquidity fragmentation while expanding the range of assets that can be used as collateral, improving the capital efficiency of perpetual contract trading or spot trading.

However, in DeFi scenarios, the only feasible way to securely implement portfolio margin is to build upon a sound and robust lending infrastructure. This is why portfolio margin has struggled to achieve large-scale success in the DeFi space—because it requires a robust lending system as its foundation.

In centralized exchanges, when a user enables portfolio margin, the exchange can essentially create a balance for the user "out of thin air" on its internal ledger. For example, if a user collateralizes BTC, the exchange can grant them a credit line to trade perpetual contracts denominated in stablecoins. This is feasible because the exchange believes the collateral is sufficient to cover this "internal credit balance."

In most cases, this model is manageable, but in the event of a black swan event, this mechanism can lead to bad debts. To be fair, centralized exchanges face an extremely complex problem, and overall, they manage it quite well. Margin trading is also an important feature they offer to their users.

However, this method of "internal credit expansion" is unacceptable in DeFi protocols. If a DeFi protocol adopts this model, it essentially assumes risk—anyone holding funds in the protocol is indirectly exposed to the risk of default on collateralized pooling. This violates the core principle of DeFi: every dollar in the system should be a fully verifiable and fully backed asset, without any implicit risk. For decentralized protocols, this is an unacceptable trade-off.

Therefore, in Hyperliquid's design, portfolio margins do not create credit out of thin air, but are fully backed by a real lending market. The lending market's pool of funds directly provides real liquidity support for lending activities under portfolio margins. This mechanism is similar to Aave V3's model, but is more deeply integrated at the architectural level.

This construction method is more complex, and users may not even perceive the underlying structural changes, but it is more robust and scalable. The key is that when market demand for margin for a certain asset suddenly increases, the system does not increase the overall platform risk. Instead, it adjusts through the interest rate mechanism—borrowing rates will rise in the short term, attracting more suppliers to enter the market to obtain profits, and ultimately, market competition will push the interest rate back to the equilibrium level. In other words, demand shocks do not lead to increased protocol risk, but rather short-term fluctuations in interest rates. This is a healthy market adjustment mechanism.

I believe Hyperlend, along with other lending protocols, will integrate into the system in a highly synergistic way. The combined margin feature will drive demand from HyperCore to the EVM layer, while HyperEVM will become the infrastructure for lending and asset tokenization. Lending activities will be completed on the EVM, simultaneously supporting HyperCore and enabling efficient distribution of liquidity both within and outside the ecosystem.

Regarding market competition, project vision, and the future of the industry.

What is the real message behind "We have no competitors"?

Kevin: Last year I saw your banner that said "We have no competitors." Many people felt that this was a bit of self-deprecation? Or perhaps, most people haven't truly understood Hyperliquid?

Jeff: I think the key to this question is—who exactly is competing with you?

I don't mean there's absolutely no competition. In a sense, there are many competitors. This is because Hyperliquid sits at the intersection of many sectors: trading, blockchain, lending, infrastructure, community ecosystem, and so on. There are many participants in the financial industry, and many people consider Hyperliquid a competitor, but I don't think that understanding is entirely accurate.

In fact, in many scenarios, it is more about collaboration than competition.

I'm slowly seeing this trend happening. Different protocols are starting to build products within the Hyperliquid ecosystem. Just a few days ago, a developer told me, "We're a multi-chain deployment team, but our structured product deployed on Hyperliquid has become our best-performing product over the past year." They were quite surprised themselves.

Similar stories will continue to emerge.

Therefore, I believe that people naturally view the market with a "zero-sum competition" mentality, but the reality is more like "competition within cooperation" rather than diluting each other's value.

On the other hand, I do think that what Hyperliquid is building is something that no one is trying to do in exactly the same way.

HyperCore's architecture and direction are largely a result of its ecosystem development path. This wasn't a grand blueprint designed from the beginning, but rather the product of over three years of continuous evolution. The journey hasn't been smooth sailing; community members have been with us since day one, experiencing many ups and downs.

The development of any ecosystem is path-dependent.

Hyperliquid has adhered to the "no insider" principle since day one, as well as fairness and openness towards users and developers. This culture has shaped its ecosystem structure today. I hope that no matter how large it grows, it will continue to uphold these values.

It is this principle that drives us to build things that others have not tried.

For example, when other teams launch perpetual contract products, few people say, "We believe that the community and developers can build an ecosystem together. Even if there is fragmentation and internal competition, it will eventually form a stronger system."

But on Hyperliquid, this has happened many times, and it will continue to happen.

Every time someone tries to build a system that can rival or even surpass centralized exchanges in a decentralized manner, someone will question: Can the community really do it?

We just keep proving this.

So when I say "we have no competitors," what I really mean is that I find it hard to find a competitor with the exact same goals and methods on the path we are trying to build: "community-centric, fairness-based, and on-chain mechanism-based challenge to centralized alternatives."

This is not to deny competition, but to illustrate that we are taking a very unique path.

The long-term costs of fairness and integrity

Kevin: You often talk about "integrity" and "fairness." It's easy to tweet that you're fair, but putting that into practice is very costly. For you, where is the biggest cost?

Jeff: This question is difficult to answer with specific examples because internally, it has never been a choice between "whether to be fair or not".

We don't sit down and discuss, "Is this fair or unfair? Should we do a cost-benefit analysis?" That's not how we think. So I can't pinpoint a specific trade-off and say that this is the price we pay to uphold fairness.

But I have no doubt that such a price must exist.

If you're willing to compromise on integrity and fairness, and take shortcuts, your growth will almost certainly be faster. This is undeniable.

Such things have indeed happened in the market. For example, FTX was a typical case of rapid expansion. They grew at an extremely fast pace, exhibiting "hyper-scale" growth. If it hadn't been exposed, they might have been able to conceal their problems for a long time.

For some, this is indeed a cost-benefit analysis.

But for us, this is not a business decision that can be calculated. Because if we compromise on the issue of integrity, the long-term costs could be "infinite." That kind of risk is not a loss that can be quantified, but a fundamental destruction of trust in the entire system and the community.

For us, integrity and fairness are not bargaining chips, but inalienable principles.

Even at a high cost, this principle itself will not change.

The goal is for Hyperliquid to become the "Internet of Finance".

Kevin: What do you hope people will remember Hyperliquid for in the future?

Jeff: I don't actually want it to be "remembered". I just want it to be used.

Ideally, it should be like Bitcoin—a long-term store of value, not controlled by any single power, and truly belonging to everyone. I'm not sure how many projects in the crypto industry will stand the test of time, but I hope Hyperliquid can become the "Internet of Finance."

Why stay in the crypto industry?

Kevin: Many founders and investors are exhausted. The crypto market has been disappointing for many over the past few years, with little apparent innovation. Meanwhile, the AI ​​field is siphoning off a lot of talent. As one of the most respected Builders in the industry, can you give some reasons—why stay in the crypto space?

Jeff: From Hyperliquid's perspective, I think humanity in our time has a very important task.

AI is a "time bomb." It's rapidly approaching and will eventually replace human intelligence on a large scale. Personally, I don't think it's as close as everyone imagines—for example, current AI can't write truly high-quality, critical code. I don't think the truly important code should be written by AI. Perhaps I'm a bit conservative, but the trend will definitely come.

Before AI can achieve "self-accelerated evolution," self-improvement, and to some extent marginalize human intelligence, we must first accomplish one thing—establish a financial system that can be accessed by machine intelligence.

When intelligence is primarily driven by machines, value transfer will inevitably also be primarily driven by machines. However, AI cannot be integrated into the existing traditional financial system. In the traditional financial system, "information" and "value" are not on the same level—code and assets are separate. That kind of system is closed, non-programmable, and lacks true openness.

Therefore, a crucial mission for us as human beings is to build a financial system in which humans still retain rights—a programmable, accessible, open, and autonomous system. Otherwise, when a new machine-dominated financial system emerges, humanity may be excluded.

In my view, Hyperliquid is our most promising attempt to date. Of course, it's not just the result of one team. Hyperliquid should be seen as the result of a collective effort—an ecosystem of primitives, protocols, companies, and builders working together.

When the true "AI era" arrives in the future, if there is a financial system that can be naturally accepted by machines, it must be a system that is well-constructed, fair enough, autonomous enough, and permissionless.