Editor's Note:
Over the past few years, the crypto industry has experienced frenzy, collapse, rebuilding, and division. Endless controversies on social media continue to surround issues related to exchanges, custody, market-making, governance, and asset issuance. The emergence of Hyperliquid has drawn widespread attention precisely because it attempts to squarely address these questions: How to build a truly usable large-scale trading market in a public, transparent on-chain environment? This article focuses on Hyperliquid and its founder Jeffrey Yan, documenting how this experiment unfolded and the price it is paying.
The following is an excerpt from the main text:
Jeff turned down a $100 million funding offer, air-dropped billions of dollars to strangers, and now can't travel without a bodyguard. This article tells the story of how he has turned blockchain and the crypto exchange Hyperliquid into the world's most profitable per capita startup.
On a Friday in January, before dawn, a 43-year-old man was taken from his home in Saint-Léger-Sous-Cholet in western France. He was driven to the small town of Basse-Goulaine, 30 miles away, where he was beaten, bound, and left abandoned. Twelve hours later, as the sun was setting in the suburbs of Paris, three men with a handgun kicked open the gate of a home in Vélizy-Villacoublay. In front of the children, they beat a couple, tied up the family of four with zip ties, ransacked the house, and then left for the train station. This was the 70th such attack worldwide in less than a year.
Two days later, I boarded a flight to Singapore.
I was visiting a team of only 11 people, but the first person I saw in their office was not part of those 11. He was a burly American man with a crew cut and stubble, sitting behind a small table in a corner of the lounge area with an Apple laptop in front of him. His build made it clear he wasn't there to code. He was a bodyguard.
One of the company's co-founders walked me from the hotel to the office. Her online alias was iliensinc, short for Aliens Incorporated. As we walked, the rain trees arched over the street, she told me they didn't use to have an office in this part of Singapore. The company originally operated out of a shared office space in the financial district, but her co-founder—the only one in the team who didn't work under a pseudonym—began to attract too much attention. At first, people just stared at him, trying to remember his face. Then strangers started approaching him. Later on, someone followed him into the elevator to his apartment. So, the company moved to a quieter place, a building where no one would think to look for them.
Even the cleaning lady had no idea what they were doing. In her eyes, she was providing services to a peripheral company that manufactured plush cat toys. Considering there were indeed 34 plush toys in the office, this misunderstanding was not hard to grasp. The company's mascot was a cat named Hypurr, with 12 of them sitting upright on a shelf. But there were also sharks, lizards, koalas, penguins, and dragons in the office, with several of them perched on Dell monitors like fuzzy gargoyles. Most of the plush toys belonged to an engineer. His wife wouldn't let him bring any more home, so he brought them to the office instead. The team never corrected the cleaning lady's misconception.
The reason was that Hyperliquid — a blockchain and cryptocurrency exchange — was one of the most profitable companies per capita globally. Last year, its 11 employees generated over $9 billion in profit. The company, founded just three years ago, now had a market capitalization of $100 billion and had never taken a cent of venture capital. The key figure behind it, Jeff, 31 this year, had almost involuntarily become one of the most recognizable faces in an industry where "the more successful, the more likely to be kidnapped" is a high-profile concern.
Prior to founding Hyperliquid, Jeff lived in Puerto Rico, single-handedly running one of the largest anonymous trading operations in the crypto space. The company was called Chameleon Trading — 'Chameleon' was his gaming handle from junior high. He initially started with his $10,000 in savings, and over the next two and a half years, the company grew at a rate of several thousand percentage points each year. When he told me his returns, he immediately tried to persuade me it wasn't that impressive. I noted his protest and another thing: Chameleon had made him extremely wealthy. At 27, he was free. To surfers, bartenders, and waitresses in San Juan, he was just a young man in board shorts.
Now, he sat barefoot in a gray armchair in a heavily guarded office in Singapore, wearing black shorts and a navy T-shirt, explaining to me why the entire financial system needed to be rebuilt from scratch. What I really wanted to know was why he switched from the first life to the second.
He said it wasn't for the money. Jeff didn't come from wealth, and his current lifestyle showed no interest in the "rich adult life." He wore the same Lululemon shorts and T-shirt every day — 15 pairs of shorts, 10 shirts, three colors each. There were no traces of wealth in the office either. The furniture was left by the previous tenant. The team only added two board games, NFTs on the wall, and those plush cats. I confirmed this when I found four books on the shelf, one of which was Frank Slootman's "Amp It Up," a management book whose core premise is that most people don't work hard enough. I brought this up to iliensinc. She shrugged. Those mantras are their own, not from a book. In the kitchen, there were three bottles of Grey Goose vodka and Macallan whisky, untouched since a community event two years ago where they failed to meet the minimum spend. This team drank tea.
Nor is it because of a love for the crypto industry. Since its early October peak, Bitcoin, the industry's most iconic asset, has fallen by about 30%. Meanwhile, gold, the asset that Bitcoin was supposed to replace as a store of value, has risen by 7% over the same three months. Most other tokens have fared even worse. When I asked Jeff how he viewed the external negativity toward the industry, he did not defend it.
“There is indeed a lot of shady behavior in this space,” he said. “Perhaps making people realize that things are not as advertised is a healthy thing.”
He does not see Hyperliquid as a crypto company.
“People wouldn't say a company is an ‘internet company’ nowadays, right?” he told me. “We use crypto tech, but it doesn't define us.”
Including Jeff, out of the team of 11, only two had previous experience in the crypto industry before founding Hyperliquid. This was somewhat intentional. According to Jeff, early adopters in the crypto space were primarily concerned with quick gains. He, on the other hand, says he is in it for the long term, aligning more with those who think more like technologists rather than traders. However, it is also a supply issue. Hyperliquid hires from the pool of international math competition medalists. Jeff won a physics gold at age 18. One of his engineers won a silver in informatics, and another was trained in the U.S. national team system. Jeff wants to bring in more people like them. In fact, since my visit earlier this year, he has brought on two more. But the pool of individuals willing to enter the crypto industry at this level has long been diluted by years of scams, lack of trust, and more recently, the AI wave.
So, since he has already made enough money to do anything, why is Jeff here?
At least from an external perspective, the answer is becoming clearer.
Hyperliquid is a blockchain that has a native exchange built on it. In a traditional exchange, a company holds your funds and controls the infrastructure. On Hyperliquid, your funds are always under your control, and the platform is public. The vision Jeff paints for it—without a hint of irony—is to carry the entire financial system. Whether this is ambitious or absurd depends on whether you are fixated on those plush cats or on the platform's numbers. Because, over the past few months since my visit, some markets that have operated in a century-old manner have begun to bend in imperceptible but quantifiable ways.
Hyperliquid started in 2023 with perpetual contracts. A perpetual contract is a derivative and also the largest single market in the crypto space. In a perpetual contract, you bet on the price of an asset you don't actually own, and unlike traditional futures, it never expires. The market formed around such bets is 6 to 8 times the size of the spot market, totaling around $7 trillion per month. Until recently, this was almost entirely handled by centralized exchanges, with Binance leading by far in terms of volume. No decentralized platform had been able to truly challenge it before. Hyperliquid was the first to do so, and its market share has now grown to about 14% of Binance's.
Then, in October 2025, Hyperliquid did something a centralized exchange couldn't do: it allowed anyone to launch new perpetual markets on the platform, as long as the asset had a price oracle. An independent team named Trade[XYZ] became the most active deployer. They initially launched the silver market. By January of the following year, its 24-hour trading volume had reached about 2% of the CME. CME, the Chicago Mercantile Exchange, is the world's largest derivatives exchange, founded in 1898. Next, Trade[XYZ] launched the crude oil contract. Crude oil has always been traded on markets closed on weekends. However, in late February on a Saturday, the US and Israel began bombing Iran. CME closed, but Hyperliquid did not. Crude oil's daily trading volume surged from $21 million to $3.7 billion. A month later, Trade[XYZ] launched the S&P 500 perpetual contract and received official authorization from S&P Dow Jones Indices. This market trades around the clock, even on weekends.
Today, the most influential products on Hyperliquid are increasingly built by those who neither work for Jeff nor will ever work for him in the future.
Trade[XYZ]'s founder requested to remain anonymous. He bought his first Bitcoin for $66 in 2013 and had been an investor for many years, not a builder. He didn't originally intend to start a company. He told me that if it weren't for Jeff, he probably would have left the crypto industry long ago.
"Hyperliquid has the opportunity to save the crypto industry," he said.
However, all this still can't explain why Hyperliquid might end up as Jeff described—in an industry where something always seems on the verge of becoming a reality but often collapses at the last minute; nor can it explain why he would give up the kind of life in Puerto Rico to validate this.
These questions lingered in my mind throughout the first afternoon at the office. It was then that iliensinc and I were sitting in the break area, with a plush cat on the table and the lingering scent of ginger and sesame from leftover lunch in the air. She told me that three years ago, when Jeff announced Chameleon's end, the team also asked him the same questions. Her answers didn't start with crypto, but with Jeff himself.
“You should ask his mother,” she said.
Jeff liked to have meetings outdoors. We were sitting on a covered terrace, where four gray lounge chairs and a coffee table were arranged. Cars passed by on the street below. Every few minutes, a gardener would start a lawnmower. The intermittent sound of a pedestrian crossing alert echoed.
Jeff tucked his feet underneath him. When I mentioned his mother, he paused for a moment.
He said, she often quoted a Chinese idiom: “There are people beyond people, and skies beyond skies.”
Essentially meaning: you may think you're strong, but there are stronger ones above you; beyond the sky you see, there are higher skies. She wasn't the kind of mother to push her child, but she wanted him to understand that, no matter how excellent he thought he was, what he saw was only a small part of the external world.
He and his sister were raised by their mother alone, living in the heart of the most lucrative stretch of geography in American business history: the Redwood Coast, located between San Francisco and Palo Alto. The mirrored glass Oracle headquarters towered over the community. Neighbors were engineers and product managers, and their children, from then on, were being raised to be the kind of person Jeff would eventually become. Jeff's parents were both Chinese immigrants, and they divorced when he was in third grade. His father left. His mother, an accountant, worked late into the night during tax season. He saw it all.
“I could feel that other families were more affluent than ours,” he said, “but I never resented it. Going out didn't cost much anyway.”
His school didn't have a strong culture of academic competition. Despite his mother often saying, “there are skies beyond skies,” she didn't push him. Until adolescence, almost no one pushed him to do anything. He went out to play, went to school, came back, then went out to play again. By the standards of his ZIP code, he was the rarest kind of child: one raised truly free-range.
In eighth grade, a friend who had just transferred from a private school invited him to join a math competition. The friend just wanted a buddy. Jeff had never seen anything like it before. Math at school was nothing like this. There were no formulas to memorize, no computational problems to mechanically work through. You'd get one question, sometimes just one sentence, and you had to find your way in. The answer wasn't a number; it was a proof—a complete argument showing why something must be true. In the end, they'd rank everyone, just like they did with sprinters. For Jeff, this was like the most fascinating part of a sport, merged with the most fascinating part of understanding the world.
That summer, he woke up at five every morning, downloaded past competition papers online, and worked on them alone in his room. He had no mentor, nor could he afford to attend a summer program. No one forced him to do any of this.
“Later on, I realized that I'm actually quite competitive,” he said. “There was this competition, and others had been running towards it since they were young, while I was lagging behind.”
A year later, in ninth grade, he had already been selected for the USA Mathematical Olympiad Training Camp, a camp comprising the top 50 high school students in the country. He was one of the youngest in this group. He didn't make it to the national team, but he said he didn't mind. During those three weeks, he sat with a group of youngsters who could stare at three sentences for five hours and extract truths that most people couldn't see at all.
Jeff told me that the math world doesn't have a Roger Federer-like universally known superstar, but at the highest levels, there is indeed some Federer-like quality. The way a proof is crafted has style and elegance, and in that training camp, he saw this up close for the first time.
“It's like being able to play football with Tom Brady,” he said, “just the geeky version of that feeling. Something most people never get to experience in their lifetime.”
The following year, he was eliminated in the mid-level selection round of a math competition. He was 16 at the time and had to wait a whole year to try again. I asked him if this was his first experience of failure.
“Losing is quite a common experience,” he said. “Most people are losers to begin with. There's usually only one winner.”
The issue was not the failure itself but the sense of emptiness that came with it.
“I felt like there was a void in my heart,” he said, “I needed to learn something new.”
So, he got hold of some physics textbooks meant for senior students. The school only offered this course in the final year, but he had just learned calculus and grasped for the first time what it was used for. He started reading the Feynman Lectures.
“I binge-read them like a TV series,” he said.
Within a year, through self-study once again, he became one of the top five young physics players in the U.S.
He made it to the U.S. Physics Olympiad national team, traveled to Estonia—his first time in Europe—and secured a silver medal. The following summer, in Copenhagen, he won a gold medal, ranking 24th globally. He was 18 then, returning to the Bay Area with a newfound understanding of “the infinite beyond”: above him, strictly speaking, there were still 23 more.
Harvard University covered almost all of his tuition. In his spring semester of freshman year, Jeff took Computer Science 124 "Data Structures and Algorithms." This course is mainly taken by sophomores and juniors and is known for being "torturous." Students in Harvard's course catalog described it as "necessary evil," with one commenting, "No social life. You are destined to be single." The course had a total of 150 students. As a freshman, Jeff ranked first in the class by a significant margin.
At Harvard, students are assigned to upperclassman dorms after their freshman year. Jeff was assigned to Pforzheimer House, where he formed a close bond with Scott Wu, who was two years his junior. The two had first met at a summer Olympiad program. Wu had represented the U.S. at the International Olympiad in Informatics for three consecutive years, achieving a perfect score in his final year, before co-founding Cognition AI. When Wu was assigned to Pforzheimer as a sophomore, he texted Jeff, "Yo, I'm also in Pfoho." Jeff replied, "That's great!"
Wu would find Jeff by the grand piano in the common room—Jeff was teaching himself jazz at the time, repeatedly playing certain phrases until they truly "clicked." They played chess, Go, poker together, and spent hours discussing what it means to truly excel at something. Jeff would talk about Faker—the greatest player in League of Legends history—as well as about Go masters and top high-frequency traders.
"He was always thinking, why is someone special," Wu told me, "What is the essence of this field? And what does truly reaching the pinnacle mean?"
In Wu's memory, Jeff was an extremely non-conformist person. Most Harvard students absorb the same information in the same environment and often come to roughly the same conclusions. Jeff never did. He was also very humorous.
"That very deadpan humor," Wu said, "He would say something completely unexpected to you, but his expression and tone were as dry as can be."
Every summer break, Jeff would work. He interned at Google X, working on the self-driving project (later became Waymo), and at the trading firm Tower Research Capital. In his senior year, he did a part-time internship at another self-driving company, Nuro, mainly because he found at least one of his four years in college to be redundant.
In his junior year winter, he and Wu together became part of the inaugural internship program of Hudson River Trading, one of the world's most successful quantitative trading firms. Among the same group of interns were Alexandr Wang and Jesse Zhang, who later founded Scale AI and Decagon, respectively. This internship program was designed as a three-week competition, and in each round, Wu and Jeff always took first and second place.
After graduating with a Bachelor's degree in Mathematics and a Master's degree in Computer Science, Jeff joined HRT full-time at the end of 2017, assigned to the US Stock Algo Team. Every week, he would have a meeting with his manager. This manager had mentored many newcomers before. Typically, these meetings followed a set rhythm: the newcomer would hit a wall in the code, the manager would help them work through it, and then the newcomer would go back to hit the next wall.
But Jeff never hit a wall. The manager recalled that he came in with his own ideas. The meetings were unusually efficient, but there was always a certain feeling that made the manager subtly uneasy. After a while, he realized what it was: Jeff seemed to do everything right, yet these accomplishments seemed to carry no weight for him personally. Eight months later, when Jeff came to tell him he was resigning, the manager immediately understood. The wording he used to announce Jeff's departure in an internal email was unusually warm, something rarely seen in the company culture.
Jeff actually loved HRT. He felt that trading was the most pure game in the real world. You were either right or wrong, and the market would tell you the answer. Many of the world's smartest people were competing with you, and in this brutal game, the results of these collisions produced a highly valuable product for the world: a market that was liquid and efficient.
However, the problem was that he had spent eight months optimizing a system that was already very good, and he was working at a company that would still be excellent even without him. This meant that he could never quite answer that lingering question:
What value have you truly added to this world?
In December 2017, the answer came knocking on his door. Bitcoin was nearing $20,000, Coinbase became the most downloaded app in the US, and billions of dollars poured into ICO projects like Jesus Coin. It was the "Crypto Christmas." Jeff first heard about Bitcoin during his internship at HRT, when two former partners came to introduce the concept to the interns. No one was impressed at the time. But later, still working at HRT, he read the Ethereum whitepaper. The paper described a globally consensus-driven computer that no single individual could shut down. He was exposed to finance every day and could see the underlying logic that powered financial systems. The whitepaper described a way to replace trust with code.
"I felt like I could go build something that fundamentally changes finance," he said.
He left HRT around April 2018 to create a prediction market where users could bet on anything with a clear outcome, such as weather, elections, or sports events. This platform would run on the blockchain, with no single entity controlling the funds. The architecture was built on an idea that Jeff believed he and his co-founder were among the first to think of: off-chain matching, on-chain settlement because Ethereum was too slow to power a real exchange. Funds would be held in smart contracts managed by code, but the user-facing interface would be fast and streamlined. It retained the promise of decentralization in the crypto world without the cumbersome friction.
He worked on this project with his college roommate, Brian Wong. Brian also left HRT. The two of them built it in San Francisco during the first Binance Labs incubation program and named it Deaux.
Kalshi, founded in 2019, adopted the same approach. Polymarket followed suit in 2020. Today, the combined valuation of Kalshi and Polymarket exceeds 40 billion dollars.
Meanwhile, Deaux only attracted 100 users.
As Jeff reached this point, the sky in Singapore suddenly opened up. The raindrops were large and heavy, filling up the drains in minutes. From the balcony, we could hear the rain pounding the streets, cars sloshing by on wet roads, tires emitting long, drawn-out hisses.
“It was never going to work,” he continued.
By the time Deaux officially launched, Bitcoin had plummeted over 80% from its peak. Jesus Coin was also long dead, beyond resurrection. No one really wanted to bet on tomorrow's weather. More importantly, Jeff and Wong had hardly considered regulatory issues at the time. Kalshi spent a full three years navigating regulatory bodies before they could finally launch their product.
When Deaux shut down, Scott Wu was one of the few people on Earth genuinely saddened by it, as he was one of the five occasional users.
Jeff returned over half of the $450,000 investment to the investors. Constrained by an HRT non-compete, he went to California's Lake Tahoe with a similarly restricted friend to ski until the end of the season. He then traveled to China, Japan, and Peru in a very frugal manner. He tried to convince me that being a tourist actually requires skill. He clearly didn’t possess that skill.
At the end of 2019, after the non-compete ended, Jeff moved to Puerto Rico. There, capital gains taxes could almost legally be reduced to zero. He went with $10,000 and a vague yet strong sense that something big was about to happen.
His partner also moved to Puerto Rico with him. They shared a one-bedroom apartment near the beach, renting for less than $2000 a month. However, the term “shared” implied some form of communal living, of which Jeff left almost no time for. He didn’t even have a monitor, so he simply took over the TV, setting up his workspace in the living room. For the first year or so, he probably allocated about 30 minutes a day to his partner, with the rest belonging to the trading algorithm constantly running on that television screen.
Jeff works at least 14 hours a day, easily surpassing 100 hours per week. He started by writing scripts in Python, connecting to various cryptocurrency exchanges, and setting up programs to trade on his behalf around the clock. He monitors these programs as they run, constantly optimizing logic, tracking data, and if the system doesn't meet expectations, he takes it down and rewrites it.
His ability to do this stems from the fact that the crypto market is open to the outside world in a way that traditional finance has never seen. In the stock market—similar to the trading he did at HRT—if you want to place a single trade on a single exchange, just that, you need to connect to 13 public exchanges within the three major data centers in New Jersey, comply with a complex set of SEC regulations called Reg NMS, and also obtain CME futures data via a microwave link from Chicago, with upfront infrastructure costs amounting to tens of millions of dollars. But in the crypto market, whether you're an HRT employee or a solo operator staring at a screen, you all connect to the same shabby, originally designed-for-web HTTP infrastructure. You just need to rent a server on Amazon's cloud.
For nearly two years, Jeff's partner had no idea what was happening on the other end of the TV. Their life seemed unchanged. Rent paid, meals eaten. She knew he was deeply involved and driven, thought he was probably doing alright, but saw no material evidence of success. It wasn't until one Friday evening in the summer of 2021 when she tried to urge him to go out to a dinner she had booked a week earlier that he refused to budge.
"You don't understand," he told her, "if I don't fix this bug right now, I'll lose $100,000."
After that evening, Jeff decided to turn this into a real company. He needed someone who could help him with everything other than coding.
Back at Harvard, there was someone in Pforzheimer House who, in his eyes, seemed to have everything in life under control simultaneously—a skill so foreign to him it bordered on superhuman. But as far as he knew, iliensinc was in Asia at the time, working as a chief of staff at a venture capital firm shuttling between Tokyo, Seoul, and Hong Kong.
When he finally reached out to her, he discovered she was actually in San Francisco. The pandemic had halted travel, turning the job that used to have her crisscrossing Asia into a series of late-night conference calls from her apartment. Jeff explained his needs to her. He didn't provide a formal job description, no title, barely even clarified what exactly she would be doing. But she had spent three years assessing entrepreneurs as an investor. Whatever Jeff described, she felt this wasn't a bet to place on someone who might fail.
The company officially has a name: Chameleon Trading. iliensinc started accompanying him to Zoom meetings with various exchange business development teams, adding a layer of professionalism to this business that in physical space was just a "one-man show above a beachfront bar in San Juan." Below those behemoth-level market makers—such as Jump Trading, Tower, HRT, and Jane Street—there exists a layer of anonymous market makers, whose size is difficult for outsiders to verify. Chameleon is the most significant among them.
By 2022, Jeff began to grow restless. By then, he had spent four years in the crypto world, accessing various markets, both centralized and decentralized, and had been deeply involved. He no longer started caring only about his gains and losses. Bitcoin had provided the world with a way to hold and transfer funds without the need for a trusted intermediary; Ethereum had provided a computer that no individual could shut down. Between the two, almost everything needed to rebuild the financial system had already been outlined. Yet, this industry had built virtually nothing concrete. The two largest exchanges—Binance and Coinbase—remained centralized. The crypto industry repeatedly resurrected things it was supposed to eliminate.
That summer, iliensinc arranged an offline team-building retreat for Chameleon at a hotel in the British countryside. By then, she had expanded Chameleon into a six-person team. Jeff gave her a Bitcoin budget. The team flew to London, visited the British Museum, and spent a few days at the country estate. The leader—being away from the screen for an extended period for the first time—did not truly seem relaxed.
Upon returning to Puerto Rico, trading continued. But Jeff told the team that they were going to start building something new. He wasn't sure what that was exactly. He had some ideas, but none truly convinced him. All he knew was that Satoshi Nakamoto's original vision for Bitcoin was quietly being buried by this industry that Satoshi himself had set in motion. And this, for someone who had made millions from this industry without building anything from those blanks, should not have had such a significant impact.
To the team, Jeff seemed like "letting some air out, only to reveal a flaw."
In November 2022, the world's third-largest cryptocurrency exchange, FTX, collapsed within nine days. It had been lending user deposits to Alameda Research—a trading firm run by the founder's girlfriend. When users requested their money back, it was no longer there. Less than six months earlier, Terra, a $500 billion crypto ecosystem, also went to zero within three days. It had attempted to construct a currency backed by the system's own logic, pegged to the US dollar, but the algorithm meant to maintain the peg accelerated its collapse instead. The industry's two largest projects ever perished within less than half a year.
Jeff had seen enough.
He told his team of six that the trading stops here. Some may disagree, but the Chameleon is done. If he's wrong, they can always go back to trading. Some in the team did indeed disagree, and some did leave. But it didn't shake his decision. There were no investors to consult, no board to convince—this was his own money, and his call. And now, a new mission has emerged.
“I was overly confident at the time, thinking that FTX would be the turning point in the downfall of centralized exchanges,” Jeff told me. “But it actually helped, because it made me determined to pursue this massive market.”
The market he was referring to is the perpetual contract.
The perpetual contract originated from an insight by economist Robert Shiller in the 1990s. Traditional futures contracts have expiration dates. At expiry, traders either take delivery of the underlying asset—such as oil, wheat, pork bellies—or close and reopen their positions, each time paying a fee. Shiller posed a seemingly obvious question: if hardly anyone trading pork belly futures actually wants pork bellies, why force the contracts to expire?
The traditional markets had viable solutions already, so there was no drive to change. In 2016, a crypto exchange called BitMEX saw an opportunity. Since then, perpetual contracts have become the dominant way of trading in the crypto market. Contracts never expire, traders can hold positions with very high leverage, often 10x, 20x their capital. The fees and liquidations it brought about have made centralized crypto exchanges some of the most profitable companies in the industry.
But by the end of 2022, there was still no truly user-friendly decentralized version in sight. The reason lay in the underlying tech. Most modern markets operate on order books. Buyers state the price they're willing to pay, sellers state the price they're willing to accept, and when the two match, a trade occurs. The more participants, the narrower the bid-ask spread typically is. From the New York Stock Exchange to Binance, it's mostly the same mechanism.
But order books handle more than just matching, they also deal with a constant deluge of price updates—traders revising quotes over and over, often changing many times before an actual trade happens. And existing blockchains are terrible at this. They're slow, expensive, and unwieldy. Each update costs money and needs confirmation. Running an order book on such a chain is like trying to run the NYSE on dial-up internet.
By the end of 2022, Jeff and the team reviewed all the blockchains other projects relied on and found none that came close to their needs. So, they made their own. Three months later, Hyperliquid had a custom blockchain robust enough to run an exchange on. Jeff then spent much of that following year on Twitter evangelizing Hyperliquid, articulating what it offered, and why it was superior to everything the industry had grown accustomed to.
And the exchange's challenge is: before it's "useful," it's entirely useless. A buyer walks into an empty market and can't find a seller. The traditional solution is to pay market makers to ensure anyone who comes here has a counterparty to match. You can pay them in cash, equity, or split tokens. Hyperliquid also had a lot of people come knocking. One of them even told iliensinc directly that his company was a "kingmaker."
"If you don't pay us, you will never rise."
They didn't pay. None of them paid.
Hyperliquid launched at the end of February 2023. Throughout March and April, users were mainly a group of NFT collectors who had never traded perpetual contracts, conducting $10 small trades and learning leverage through simulated trading competitions. There were no real "Jeff Soo users" at all.
By May, Jeff took those strategies that made Chameleon one of the most successful anonymous trading businesses in the crypto space and packed them into an on-chain treasury: HLP (Hyperliquidity Provider). You can deposit $10 or $10 million. There are no management fees or performance fees. This treasury runs automated strategies, and every dollar in profit goes to the people who put money in. All accounts are recorded directly on the blockchain. If FTX were built this way, the holes in Alameda should be clear to the whole world.
HLP instantly solved two problems. It provided liquidity for trading; and the users providing liquidity also encountered something that traditional finance had never truly opened to the average person before. An early user told me that this was almost the first time in history an ordinary person could invest in high-frequency trading strategies at zero cost.
"If Jeff were willing to take a 2% management fee and a 50% performance fee, I would also invest without hesitation," the user told me, "but in reality, someone without a background, connections, sitting in any corner of the world, can access one of the greatest market-making strategies in crypto. People still don't realize how special this is."
Hardly anyone understood at the time. By autumn, as crypto assets rose every day, HLP users watched the treasury balance decline while Bitcoin surged. The algorithm itself was performing well and making money through trades, but since everything was running on-chain, it couldn't hedge against the overall market exposure. Traditional market makers can offset risks on other exchanges, something that HLP couldn't do by design. So, even though it was winning on every trade, it was essentially in a "short" position in a continuously rising market. People were furious. Other projects besieged Hyperliquid on Twitter and Discord, and Jeff fought back. It was early enough that he still took these things very personally at the time.
But HLP was never meant to be the final answer. Jeff built it with the intention of propping up liquidity before the arrival of independent market makers. He knew that those market makers would eventually see the opportunity. The demand far exceeded the supply, and the significant bid-ask spread meant that easy money was available for anyone willing to place orders. He wrote documentation, sent out lengthy tweets explaining what market-making was all about, and personally guided companies to integrate. However, most people still hesitated. Other exchanges paid market makers, but Jeff refused to do so. At the same time, HLP itself could not infinitely scale to fill this gap.
“Alameda is key to FTX’s operation,” he said. “We don’t want HLP to become key to Hyperliquid’s operation.”
Various metrics were growing, as were the complaints. In theory, the market makers should have arrived by now. But if they hadn’t come and the users left first, everything would come to an end.
However, there was always one type of people who would show up on time.
Venture capitalists.
Their analysts had been using this exchange privately for some time, going back to their partners one by one to say: this thing is really good. Consequently, the partners started making phone calls. Jeff and iliensinc had never engaged in any fundraising promotion. They didn’t have a pitch deck. The protocol was indeed generating fee revenue, but from the beginning, Jeff insisted that none of this revenue should go to the team. When VCs came online for meetings and asked for a deck, Jeff and iliensinc just chatted with them. By the end of the conversation, the other party would finally understand: no, there isn’t one.
By January 2024, some funds had already visited in person. iliensinc was familiar with the routine. Having been an investor herself, she began explaining various financing terms to Jeff and reminding him of the rights that needed special attention. For about two weeks, he went along with this process.
“It was almost like it was instinctual,” he told me. “When VCs started showing up, I thought, ‘Oh, it looks like it’s time to raise.’”
His only condition was that he would only consider a term sheet with a valuation of over $1 billion. This was less than a year since Hyperliquid went online. The team was burning through hundreds of thousands of dollars every month, all coming from Jeff’s savings. When an investor named the price he wanted, Jeff spent a weekend thinking it over carefully.
He went to those who had started companies before and asked the VCs themselves: what was the true meaning of raising funds? However, they couldn’t convince him why “their money” would be more valuable than “the money itself.” At some point, he felt that saying “no” was the right decision. Once this feeling settled in, it was all over.
On Monday morning, he said to iliensinc:
「We are not taking this money.」
「What the heck?」
She could hardly believe it. She was the one handling the money, watching it burn away bit by bit. Now a family fund was willing to invest nearly $100 million, and after she had spent a full two weeks preparing for the funding, he suddenly decided to reject it. The rest of the team's reaction wasn't any better.
He called the fund, officially declining. They didn't believe it either. Did you accept someone else's term sheet? No. Hyperliquid isn't a company, it's a protocol. From day one, its neutrality is the most core thing.
「If Bitcoin had raised VC money in its early days,」 he said, 「I really don't think it would still be Bitcoin. Its most core value proposition would have been completely destroyed.」
Plus, he didn't need the money anyway. To this day, a lot of the team's expenses are still covered by Jeff himself.
On January 28, 2024, he tweeted four lines:
No investors.
No paid market maker.
Development team does not take fees.
No insiders.
Hyperliquid has only one meeting every day: the morning stand-up. On my second day in Singapore, I observed this meeting. The team gathered around a screen with an engineer. Above the screen sat a plush dragon. They were testing a new feature called portfolio margin at the time, and the discussion was almost entirely about "where things could go wrong." But for a long time, it wasn't really a conversation. Jeff would cross his arms, bow his head, and stare at his own bare feet in thought. The engineer next to him did the same. This silence wasn't awkward or brief, and no one in the room found it odd.
This meeting atmosphere is partly due to personality. The team is very young, ranging from 24 to 31 years old, and almost everyone is extremely smart and introverted. But when I asked Jeff later if he usually reads a lot, his answer made me feel like things weren't just "shy."
「By the standards of what most people consider optimal, I read much less,」 he pushed his dark-framed glasses and smiled, 「If a book really is to enter you in a way that permanently changes you, that actually takes time. In terms of return on time, it's not that cost-effective.」
He tilted his chin — a gesture I would later come to recognize — like someone trying to equalize ear pressure on a plane. One particular risk of writing about young techies is that sooner or later they’ll tell you: they don’t read. So when Jeff went on to mention that he probably reads a book every two months and also looks forward to one day sitting down to read all the books he hasn’t read yet, I actually felt relieved. Then he went on to explain why reading more books right now had to take a back seat.
“If you’re not the first one to do something,” he said, “then it’s probably not worth your time. That’s really how I see it. Acting on that premise, reading books doesn’t really help. Because if what you’re doing already has enough background material available, then it’s most likely been done before. If it’s already been done, why do it?”
By the end of 2023, Hyperliquid encountered yet another issue in the well-trodden path of the crypto world. And just like before, Jeff had no interest in playing out the scenario according to the same old script.
A token for a crypto project means that the holder can share in the project’s success. Who gets the token first and under what conditions is usually determined through a “point system.” The project team announced that user activity on the platform would earn points, and everyone assumed these points would eventually be exchangeable for tokens. As a result, hordes of people rushed in, trying to hoard as many points as possible before the exchange took place.
The issue was that most of the people rushing in were not really “users” at all. They were professional rug pull teams who would reverse-engineer the points formula, use automated strategies to maximize their returns, and then leave. The genuine users who were meant to be rewarded could only pick up the scraps.
Hyperliquid’s version went live on November 1, 2023. Users would trade on the platform and accumulate points weekly, but the plan had no public formula. No one knew exactly how it worked. Every Friday, iliensinc would announce the points for the week, and a sort of ritual emerged: users would stare at Discord, wait for her account to start typing, then all rush out to compare how many points they received, share screenshots, and speculate on how the entire system calculated the points.
“Rewarding real users is crucial,” Jeff said, “It’s hard to precisely define, but Hyperliquid’s point system may have reduced the share of ‘professional farmers’ from 99% to 20%.”
Around this time, those market makers that Jeff had refused to pay directly finally started to enter the scene one by one. One of them was one of the largest market makers on Binance. Following FTX, he was extra cautious with all new exchanges. But he had a few mutual contacts who spoke highly of Jeff. In September 2023, at a conference in Singapore, he met Jeff and iliensinc for the first time.
“Jeff is very ambitious, but not arrogant,” the market maker told me. “He was very restrained in describing what he wanted to do, and almost all the key points aligned.”
He messaged the team as soon as he stepped out the door: we should integrate.
Two weeks later, they went live officially.
And when this market maker truly integrated, he found many thoughtful features buried in the platform that only traders would notice. Hyperliquid built an “speed bump” that made it harder for the most aggressive quant firms to sniper other market makers. Later, this mechanism was widely replicated throughout the industry. The effect was that market makers could provide deeper liquidity without having to push themselves to the forefront of latency competition to survive. Jeff was essentially sacrificing a portion of the exchange’s volume — the volume generated by market makers sniping each other — to give regular users a better price.
This trade-off would reduce Hyperliquid’s own revenue.
It was also at that conference — Token2049 — that Jeff and iliensinc decided to relocate the team. Jeff told me that the regulatory outlook in the U.S. for crypto derivatives was too uncertain, and continuing to build there felt like an unnecessary risk. A lawyer I interviewed described that period as one where U.S. regulators “have almost exhausted all possible means to push this technology out of the U.S.” iliensinc had to decide between Hong Kong, Switzerland, and Singapore, ultimately settling on Singapore. It was modern, safe, and had fewer distractions.
By the spring of 2024, the team had all moved over. Jeff liked it here because the city-state was “boring” enough. His life had only two modes: work and workout. He swims, he runs, anything that gets him to the point of exhaustion without risking injury is fair game. This principle stemmed from a lightweight motorcycle accident he had in Puerto Rico, leaving a scar on his face and keeping him away from the keyboard for a whole week. The presence of workouts is just to clear his mind so he can go back to building. His only leisure concession is Sunday mornings. The rest of the week belongs to Hyperliquid. He even cuts his own hair because going to the barber takes time after all.
He doesn’t see anything abnormal about this. Or more accurately, he finds most people’s approach to work abnormally lax.
“I think people, in general, are a bit too soft,” he said. “The brain is an organ too. If you need to work longer, it can be trained.”
He had already learned not to impose this set of requirements on the team. Every day at noon, everyone would eat together, sitting around a black wooden table like a family. Every Thursday, they would eat Chipotle. Since there was no Chipotle in Singapore, they gave the menu recipe to the chef, who now prepared it. Lunchtime conversations would usually drift to what everyone was currently watching or listening to. At these times, Jeff would often fall silent, wearing a look that suggested he was thinking about something else—and most likely he was indeed thinking about something else.
It was also that spring when Hyperliquid's perpetual contract daily trading volume exceeded $1 billion, and the underlying infrastructure began to creak under pressure. One afternoon, the alarm system went off, and it just wouldn't stop. The platform couldn't handle the influx of new users. This was Hyperliquid's first crash. But what people outside the office truly cared about was the upcoming Hyperliquid token.
In May, Jeff posted a six-month roadmap on Twitter. It was filled with various technical ambitions. Not a single word mentioned the token.
In the preceding months, Hyperliquid had expanded from derivatives to spot trading. The first spot token listed was Purr, named after that cat. Launching spot was a necessary step: to issue the Hyperliquid token, the team needed a spot market to trade it. However, this also introduced a problem that a perpetual exchange had never faced. When trading perpetual contracts, no one actually needs to hold the underlying asset; you are just betting on the price. But with spot trading, someone has to custody the assets. And this was precisely what Jeff did not want to do. The whole point of the system was for users to control their assets themselves.
To address this issue without acting as a custodian, he realized he had to stop seeing Hyperliquid as a "decentralized exchange built on a blockchain" and instead understand it as a "blockchain with exchange functionality built in." The chain the team had previously built to run the exchange was already capable of processing hundreds of thousands of orders per second; if they made it programmable, it would become an open system: anyone could write code on it, build financial apps, just as thousands of developers had long done on Ethereum. The difference was that Ethereum was too slow to run a truly viable exchange, which was why Jeff had initially wanted to build his own chain.
By opening up this chain, assets could be introduced to Hyperliquid via protocol-secured decentralized cross-chain bridges without the need for any single party to custody them. And anyone building apps on this programmable layer could directly access the exchange's order book and all the liquidity already accumulated there. Developers could create lending platforms, stablecoins, mobile trading apps, and plug directly into the same market—where institutional players were quoting liquidity in the billions of dollars every day.
Jeff doesn't like analogies. He would tell you that Hyperliquid doesn't have a true analog in traditional finance; people always try to shoehorn new things into old categories rather than understand them for what they are, which is a mistake. But for those of us who are not Jeff, it's like Amazon first building cloud services to support its e-commerce platform, only to realize later that the cloud service itself might be even bigger than the marketplace. The wording Jeff first used on that Twitter was: Hyperliquid will carry the entire financial system.
He didn't actually want to take this step. He told me that subconsciously he never wanted to give himself another task. Integrating the virtual machine into Hyperliquid is a massive undertaking, and the team had no idea if it was feasible or how much work would have to be done from scratch. But at some point, it became too obvious: if they didn't do it, they would spend the next few years cobbling together some "kind of like Binance, kind of like Ethereum" components, neither of which would work, and they would regret it in the end.
The community was furious. Everyone was expecting an airdrop, but instead, they got a tweet discussing infrastructure upgrades. Many of the top comments were quoting lines from "Breaking Bad": "We had a good thing going." "I hate this. You've betrayed us." Users don't want a blockchain; they want money. Xulian— the person who joined the team due to a user interview that was supposed to last only 15 minutes but dragged on for an hour and a half— bore much of the brunt of the anger.
"Jeff was thinking long-term optimal," he told me. "We really don't care if something doesn't look good immediately."
According to iliensinc, the loudest complainers eventually got tired. Over the next six months, the team focused on advancing the spot market, building the programmable layer, testing on a separate network, and preparing for staking. Then, on November 29, a Friday, the HYPE finally arrived.
Hyperliquid airdropped 31% of the total token supply to approximately 94,000 early users. No strings attached, no unlocking period. If you had ever used the platform and earned points, that morning you woke up, the tokens were already in your wallet, richer than the night before. By the opening price, this airdrop was worth over $1 billion; at its all-time high, it briefly hit $16 billion. It was the largest wealth transfer in cryptocurrency history, and every dollar of it went to the users.
The team's own allocation is 23.8%, even smaller than the community share, and is vested over multiple years. On the airdrop day, they didn't receive a penny. The VCs also received nothing. If they wanted to hold this token, they could only buy it on the open market at the same price as everyone else, and the only place to buy was Hyperliquid because the token was not listed on any other exchange. Listing elsewhere would also incur costs; they didn't spend any.
That morning, Jeff didn't need to explain much on Twitter.
"Waking up to find I received a mid-six-figure airdrop." one user wrote.
Another replied: "Today, HYPE changed my life. This money is enough to live comfortably for many years in the future, help my family, and continue to participate in the bull market with a heavy position."
Someone else wrote: "Received a seven-figure airdrop, thanks Jeff, divine blessing."
"I feel great," Jeff told me, "In most cases, those who truly support something early on often cannot share in its later upside and do not get meaningful ownership. But this time is different."
I asked him how he felt since everything had been given such a public price tag.
"Terrible," he said.
It was a Wednesday night in late March 2025 when iliensinc's computer suddenly started blaring an alarm. She was on the phone at the time, so she hung up immediately. On the screen, the balance of HLP, the Hyperliquid community treasury, was plummeting.
In the days leading up to this, a trader had been probing Hyperliquid's defenses with small, coordinated positions. Now, the probing was over. The other person had opened three positions on an obscure token called JellyJelly. This token had a total market cap of about $15 million and a daily volume of only $72,000. One was a large short with the other two being longs. The short position was designed to blow up. This person was first shorting a token they knew was about to skyrocket, and once the position collapsed, they'd pass this hot potato on to someone else. Like pulling the pin on a grenade and then passing it to someone else.
That "someone else" was HLP. On Hyperliquid, if the order book couldn't absorb a trader's liquidation fast enough, the community treasury would take over that position and gradually unwind it later. Normally, this was a routine mechanism. However, JellyJelly had minimal order book depth, so once HLP was stuck in it, there was no way out. Meanwhile, the trader was aggressively buying up JellyJelly on the open market. In less than an hour, the price surged over 500%. With each price jump, the treasury's losses increased.
iliensinc stared at the screen, watching the loss pass $5 million, $8 million, $12 million. There was no mechanism in the system to stop this. No one had ever designed a world where someone would wield a $15 million token as a weapon.
Across Asia and Europe, validators began to come online. Hyperliquid's blockchain was maintained by over twenty independent validators, responsible for validating every transaction and securing voting power by staking large amounts of HYPE. Many had been using Hyperliquid long before the HYPE token existed. They, like anyone anywhere in the world, could see what was happening on the same public ledger. But this was no ordinary transaction to them. Within minutes, all validators voted to delist JellyJelly and settle positions at pre-man
