Guest: Haseeb Qureshi (Managing Partner of Dragonfly, former professional poker player and senior software engineer)
Host: Kevin Follonier (host of When Shift Happens)
Podcast Source: When Shift Happens
Original title: $4 Billion Fund Manager Reveals The 5 Crypto Coins He’s Going All In On In 2027 | E194
Air date: November 15, 2026
Compiled by: Shenchao TechFlow
Original link: https://www.techflowpost.com/article/35439
Disclaimer: This article is reprinted content. Readers can obtain more information through the original link. If the author has any objections to the reprinted form, please contact us and we will make modifications according to the author's requirements. Reprinting is only for information sharing and does not constitute any investment advice and does not represent Wu Shuo’s views and positions.
Haseeb Qureshi, managing partner of Dragonfly, a top crypto venture capital firm with over $4 billion under management, was a guest on the When Shift Happens podcast and had a 70-minute ultra-in-depth panoramic conversation with host Kevin Follonier. As a senior investor who has led and witnessed the growth of more than 150 crypto unicorns, Haseeb reviewed in detail for the first time the most profound lessons and highlight battles in his investment career, including the "workbook-style" bias that rejected Solana's seed round at a valuation of 4 cents in the early years, and the underlying mathematical logic of leading the investment in Ethena with a heavy position in the industry after the collapse of Terra.
Facing the AI wave that is reshaping the global technology landscape, Haseeb sharply pointed out that there are striking structural similarities between the current AI investment craze and the crypto summer of 2020. He not only dissected in detail the deep revelation that Dragonfly was rejected when investing in Hyperliquid at a valuation of US$600 million, but also completely outlined the top investment panorama for 2027: why the privacy AI platform Venice is the position with the strongest conviction at the moment, and why Zcash is facing a cryptographic catastrophe and compliance penetration caused by AI's rapid problem-solving constitutes Bitcoin’s indispensable ultimate hedge, and why Silicon Valley’s culture of embracing failure and trying new things is the only antidote to system rigidity.
Summary of key points
Wonderful views
text
Fanaticism and Illusion: Why is the current San Francisco AI investment like the crypto summer of 2020?
Kevin Follonier: A party is obviously not the best place to talk about serious topics, but for some reason, people in the crypto community always seem to be extremely enthusiastic about throwing various parties.
Haseeb Qureshi: This is definitely not exclusive to the crypto community. Friends, if you go to San Francisco now and go to those gatherings and parties in the AI community, the scene is exactly the same. The scene that played out in the encryption industry three or four years ago is now being completely repeated in the AI investment circle.
Haseeb Qureshi: It’s a completely socially driven investing frenzy, with barely any serious due diligence to speak of. Term sheets were written directly on napkins, checks were flying everywhere, and valuation metrics were completely out of whack. You often see so-called "cutting-edge labs" where just four or five young guys and a conceptual idea can make crazy money at ridiculous valuations of billions of dollars. History is always surprisingly similar, everything old becomes new. We have experienced the entire script before.
Kevin Follonier: What struck you most when you saw this scene with your own eyes?
Haseeb Qureshi: This is purely human nature. Whenever the public is extremely excited about something and feels that "they must participate immediately, otherwise they will be abandoned by the times," investment principles will succumb to the surrounding enthusiasm and social atmosphere extremely quickly.
Haseeb Qureshi: Looking back now, it’s easy to laugh at our own fanaticism. Everyone recalls 2021 and 2022 with a hint of self-deprecation: We thought we could take over the world, and felt that the Bored Apes represented the future of human culture. People could just throw money into any ridiculous and stupid project, and then easily get 3 times, 5 times or even 10 times huge profits. Everyone fantasized about being rich forever as long as they did nothing.
Haseeb Qureshi: And if you go to San Francisco today and talk to those young entrepreneurs, you will find the same obviousness in their eyes: "Of course I should be rich! I don't even need to make any commercial product, I just need to be in the right place at the right time. Of course I can get a huge check from a top venture capital investor, and then naturally cash out my secondary stock (Secondary) in a high valuation round and leave."
Haseeb Qureshi: I still remember that one of the most incredible things to me during the crypto craze in 2021 was that entrepreneurial teams began to openly resist equity maturity (Vesting). Today in San Francisco, this has actually become the norm! People are discussing on social media and in cafes with confidence: "Why do we have to have a one-year maturity cliff? Requiring employees to stay for a full year before they can get equity seriously hinders recruitment! What if artificial general intelligence (AGI) comes in six months? The opportunity cost of staying for one year is too high. Who would be stupid enough to sign an option agreement with Cliff?"
Haseeb Qureshi: It turns out that OpenAI really eliminated the option maturity cliff period internally! The direct reason for their cancellation is that all the AI startups outside have already canceled Cliff in order to grab people. You often hear stories about someone joining an AI lab as a co-founder, leaving after three months, then starting another lab, staying for a year, and then leaving again. These plots were exactly the same in the encryption circle back then. This is the irresistible instinct of human beings when faced with the speculative frenzy.
Kevin Follonier: Now that you've seen reruns of the movie several times, how much of an urge did you have to be involved in it as you were watching this happen? Do you think this is 100% pure bubble, or does it really have a revolutionary core?
Haseeb Qureshi: The current craze is indeed extremely crazy, but there is indeed a very solid reason behind why it can happen: a real technological revolution is taking place for real.
Haseeb Qureshi: This is just like the crypto market from 2020 to 2021: although the market at that time was filled with a lot of scams and garbage, 2021 witnessed the collapse of Terra, the thunder of FTX, and the ebb and flow of Bored Ape and OpenSea. There are too many things that have simply not stood the test of time; but at the same time, 2020 to 2021 The year was also the golden age when decentralized finance (DeFi) was truly born. It was during that period that the on-chain scale of stablecoins achieved an explosive vertical climb, and Bitcoin broke through the historical mark of US$20,000 in one fell swoop. All this reality and illusion are intertwined at the same moment.
Haseeb Qureshi: The same is true for today’s AI. There are a large number of pseudo-needs that are destined to fail and cannot reach any end point, but they must also contain great achievements that can completely change human society.
Kevin Follonier: So according to cycle comparison, which stage of the encryption cycle do you think the current AI is roughly equivalent to? Is it equivalent to the crypto peak of 2021?
Haseeb Qureshi: I think it's probably closer to the early mid-summer phase of 2020.
Kevin Follonier: Does that mean it’s going to get even crazier?
Haseeb Qureshi: It's going to have to go through a peak that's far crazier than it is now, before it really goes downhill and frustrates everyone. But at some point in time, tragedy is bound to happen. You only need to look at the current huge scale of global data center construction to understand: in order to repay the huge debts borne to support the construction of these computing infrastructures, the industry must create extremely astronomical net profits in the future to balance the accounts.
Haseeb Qureshi: Look at the computing power procurement commitments signed by cutting-edge model manufacturers like Anthropic, which are as high as $500 billion! Once the future macroeconomics experiences a cyclical trough, or society’s actual structural demand for AI computing power encounters any slight stagnation, mathematically speaking, it will be extremely difficult for them to fill such a terrifying capital black hole.
Investing Paradigms Through Cycles: Searching for Non-Consensus and Archeology in the Herd Mania 2021 Crash Memo
Kevin Follonier: So from your personal practical perspective, in the face of this cyclical law, is your strategy to go along with the flow and add fuel to the fire, or to be restrained and wait for the bubble to burst and everyone to be severely punished by the market before going to pick up bargains in the ruins? How do you usually play your cards at this stage?
Haseeb Qureshi: The core fact is: No matter which stage of the cycle we are in, there are almost always excellent investment opportunities in the market. At the high points of 2020 and 2021, there are certainly a large number of low-quality assets that are destined to return to zero, but if you look back, for example, Ethena’s seed round was planned by us in the first half of 2022, before the violent market turmoil and the collapse of FTX; including the high-performance public chain Monad, and the full-asset derivative Variational, we wrote the seed round check for it near the high point of 2020 or 2021.
Haseeb Qureshi: So the answer is: there is gold anywhere in the cycle. The key is not whether the outside world is a bull market or a bear market, but how much money you invest at what stage, and what quality assets you ultimately buy.
Haseeb Qureshi: The most fatal mistake in investing often occurs when you rush to invest out of fear of missing out (FOMO), or invest in a project that you don’t really understand at all. This is the trap I see most often in the frenzied market: people see Axie Infinity skyrocketing, see all the top VCs shouting that they must invest in OpenSea, and a huge amount of hot money is rushing to grab Yuga Labs. The true mentality of ordinary investors is often: "I actually don't understand why this thing is worth so much money, but those top institutions are all investing in it, and they must have some inside information that I don't know." 』
Kevin Follonier: Was this superstition that "they must have the inside story" later proven to be true?
Haseeb Qureshi: Total self-deception. After experiencing the great crash of 2021 and 2022, today we have the opportunity to conduct in-depth "ruin archeology" and consult among the debris a large number of investment memos written by institutions at the peak of the mania. When I watched them do deals to grab market shares with valuations of billions, I once thought that these first-tier funds must have done a lot of rigorous financial modeling, detailed due diligence and in-depth strategic deductions.
Haseeb Qureshi: But when you actually open those memos today, you will feel extremely embarrassed: those investment reasons are so simple and childish. They can be summed up in one sentence, "Because other top institutions are investing, so we must also invest." This herd mentality is extremely contagious during the mania period, and is also the root cause of the eventual elimination of most capital.
Missing 4 cents Solana’s profound confession: the “homework-style” fault-finding trap of junior VCs
Kevin Follonier: Now that we’re talking about investment memos, is there any memo in your fund or in your personal investment career that you feel is very embarrassing when you look back now, or that taught you a painful lesson?
Haseeb Qureshi: That was definitely the Pass Memo I wrote about Solana in my early days when I was starting out in the industry. This was in 2017-2018, my first year or so in the venture capital industry. Today, as a mature institution, we only write an investment memorandum when we decide to invest in a certain project; but as a fledgling young investor at that time, in order to force myself to learn and record the thinking of each decision-making node, I would write an extremely detailed Pass Memo for every project that was rejected by me.
Haseeb Qureshi: 2018 was the ancient era when the last round of public chain circuits sprouted. I wrote veto memos for almost all the projects in that batch, and the one that slapped me the most and made me most ashamed was Solana’s seed round. If I remember correctly, Solana’s seed round token price at the time was around 4 cents.
Kevin Follonier: 4 cents! What was the reason for the rejection you gave in the memo?
Haseeb Qureshi: If you look at the memo I wrote at that time, you will find that it is full of extremely hard-core micro-technical criticism. Solana’s business plan at the time stated in black and white that they would achieve an ultimate throughput of 600,000 TPS through GPU hardware acceleration and a new architecture.
Haseeb Qureshi: Because I am a software engineer myself, I almost regarded that financing plan as a difficult homework question from the computer department of the university! I dismantled their mathematical formulas and engineering deductions one by one, and listed several pages of technical objections in the memo: proving that 600,000 TPS cannot be achieved with consumer-grade hardware and existing network delays. I singled out their micro-bottlenecks in concurrent communication, memory bandwidth, and node verification. Finally, I triumphantly concluded: "The project's technical indicators are seriously exaggerated and are not feasible for engineering implementation. It is recommended to give up." 』
Haseeb Qureshi: This is so typical. It is the epitome of the arrogant mentality of a junior investor: you take the grand vision described by the entrepreneurial team in an extremely literal way and spend all your energy on proving that they "can't do it technically yet." Just like a teaching assistant who is eager to prove his IQ, he corrects the homework to perfection, picks out all the mistakes and puts a big red cross.
Haseeb Qureshi: I often make a joke: As an investor, the worst outcome is that you are absolutely right in all the fussy details, only to miss out on a huge return of 10,000 times! In college mathematics classes, if you write down the steps to solve a problem very beautifully, even if the final answer is wrong, the professor will still give you points for the steps; but in the world of venture capital, if the process is all correct and the conclusion is wrong, the score is strictly zero! No matter how sane your analysis is and how much you look like a genius, if you get the end result wrong, you are simply wrong.
Kevin Follonier: That’s the most fascinating thing about business investing and entrepreneurship. I still remember when I was in school, if you were a very competitive person, you would take chances when doing math problems: even if the answer was wrong, maybe the teacher would find that my previous steps were correct and give me a passing grade; but in English literature or French classes, as long as the teacher doesn't like you, he will give you a bad low score no matter how you write, which is completely subjective and arbitrary.
Kevin Follonier: I was always eager to get the best results. I was always thinking: How can I prove that I am better than others? The answer is that you must become an entrepreneur and investor! Because in front of the end of business and the market, there is only one pure standard: Is your business successful? Did your investment make any money? No marking teacher can arbitrarily give you a failing grade because of your appearance, the glasses you wear, or the way you speak. The market is absolutely objective.
Venture Capital’s Ladder Promotion: From Deal Flow, Analysis, Case Selection to Really “Winning Deals”
Haseeb Qureshi: Exactly. There is no politicking, no ambiguity. The market does not care about your family background, social circle or appearance. Right is right and wrong is wrong. The objective world will tell you in the fastest and coldest way whether you are a good investor or a bad loser.
Haseeb Qureshi: But at the same time, venture capital (VC) has extremely special and essential differences compared with hedge funds (Hedge Fund). In the field of secondary hedge funds, the trading tools faced by everyone are completely equal. If you want to buy Tesla stocks, or SK Hynix, or Micron Technology, as long as you click to place an order in the trading software, everyone can buy at a fair price. This is a truly equal playing field.
Haseeb Qureshi: But venture capital is not like that at all. In the VC industry, you can sit in the audience all day long and point out the company. With hindsight, you can say: "I knew that company could succeed. This is so easy to understand!" Sequoia actually invested in FTX, and if I were there I could definitely tell it was a scam. 』But this is meaningless, because the biggest threshold for VC is not to understand it at all, but to obtain the right to invest from the beginning (Access).
Kevin Follonier: So getting the ticket is a more central deciding factor than picking the right horse?
Haseeb Qureshi: You cannot choose between the two: the first step is to get the ticket (Access), and the second step is to bet on the real winner. Within Dragonfly, I often draw a capability ladder diagram for new junior investors, telling them how a person starting from the bottom will be assessed and promoted in the industry in the future.
Haseeb Qureshi: At the bottom of this skill ladder, the first level of ability is deal flow: can you find cases through massive communication? Can you figure out what is happening in the industry? The second level of capability is Analysis: Can you do in-depth due diligence? Can you write a solid memo that completely breaks down the upside and downside of the business model?
Haseeb Qureshi: The third level of ability is case picking: identifying who will win and who will die in many cases. But what I want to emphasize is that the ability to select projects is only halfway up the ladder! In the entire core competency system of venture capital, its weight is actually far less high than the outside world imagines, because there are actually many people in the industry who can see good projects.
Haseeb Qureshi: Above project selection, the highest and most decisive fourth level ability is called "Winning the Deal"!
Kevin Follonier: What exactly does winning a deal mean?
Haseeb Qureshi: Just take the share! There are so many people in the industry who can see that OpenAI will become a giant. Sequoia can see it, a16z can see it, and Benchmark can also see it. Everyone knows it will win, but why should you leave your precious investment share? Just because you can see a good project doesn't mean you can sit at the negotiating table. The real ranking lies in whether you can win the heart of the founder and seal the deal.
The unfamous Underdog’s battle to break the situation: borrowing the halo of a major manufacturer vs. building an independent personal brand
Kevin Follonier: Unless you have an unparalleled super brand. Just like the traditional consulting industry, McKinsey can dominate everything and charge high fees by relying on its golden name. In the field of venture capital, brands also carry great weight. If you are not from a well-known family, as a non-traditional giant or an emerging challenger (Underdog), how on earth do you win these top-level transactions in which competition is fierce?
Haseeb Qureshi: This is the most fascinating core proposition of the venture capital industry. If you are in a leading organization with a huge platform like Sequoia or a16z, you are essentially "renting" the platform's brand assets. Legends such as Marc Andreessen, Ben Horowitz and Chris Dixon have spent decades building a strong reputation. As a young investor on the platform, you are borrowing their aura to talk about deals. It is not your personal asset.
Haseeb Qureshi: But if you are not in these wealthy platforms, you have to build your brand on your own. There are many different tactical paths to winning a deal, and it all depends on the size of your check, what your appeal is, and what the opportunity cost is for founders to accept your capital. If you are a very large fund that manages huge funds, then your competitor is a16z, and you must defeat them on the frontal battlefield; but if you are a small and medium-sized boutique fund, what you replace is often just the shares of a few angel investors or strategic investors, and the resistance to your entry will be much smaller at this time.
Haseeb Qureshi: But no matter what, you must bring unique scarcity value to the founder, or make the founder truly like you from the bottom of his heart and be willing to take the initiative to make difficult trade-offs in order to accept you.
Haseeb Qureshi: There are usually two factions in terms of playing styles: one is the "heat-seeking oriented", who focus on the most popular cases in the market, and try their best to squeeze into the negotiation table; the other is the "truffle hound". They know that they cannot defeat the big giants in the frontal battlefield, so they choose to go deep into the jungle to dig for unpopular early pearls that the giants have not even noticed. Truffle Hunter's strategy, while extremely difficult to scale, often yields stunning returns for smaller funds.
Haseeb Qureshi: But if you want to achieve everlasting success in the VC field, the core moat is still to build a strong personal public brand and industry legend. Why do great founders struggle to get a check from Marc Andreessen, Vinod Khosla or Naval Ravikant? Because they have established an irreplaceable ideological reputation in the industry. Entrepreneurs will say: "Even though Vinod has a weird and picky personality, he is indeed the leader who can penetrate the fog." 』
Kevin Follonier: So this is the underlying reason why you frequently participate in podcasts and express your opinions? Building a personal brand is the ultimate weapon in any field.
Haseeb Qureshi: Yes, venture capital is ultimately a business about people. Many people think that a16z’s brand will be able to crush medium-sized institutions indiscriminately. However, in many top-level transactions with extremely fierce competition, the final decision is often not the cold institution logo, but the specific partners themselves. Because in the next five to ten years, it will be that concrete, living person who sits on the founder's board of directors, answers calls for help at two o'clock in the morning, and helps founders deduce life-and-death problems.
Haseeb Qureshi: There are often some old-fashioned tactics in the traditional financial circle, such as arranging extremely luxurious private dinners for founders, chartering private jets for founders, or giving away a VIP pass to the F1 Grand Prix paddock, trying to stun founders with these flashy ostentations. Some people may use this trick in traditional investment banks or some non-technical fields, but for the truly top technology geeks, this trick is completely ineffective.
Haseeb Qureshi: In crypto and AI, true hardcore founders only value two things: deep intellectual understanding and absolute trust. What they need is an ally who can truly resonate with them in terms of technical code, network mechanisms and crisis public relations, not a banker who can only wave checks and display wealth.
Refocusing on the causes and consequences of NEAR: Transformer paper authors, paper drafts and one-day refactoring protocol
Kevin Follonier: Let’s talk about some of the highlight cases that you dug up very early on. Let's start with NEAR, you got involved with them very early. When did you invest in NEAR specifically, and what was the decision-making logic behind it?
Haseeb Qureshi: It was way back in 2018, and I co-led a seed round for NEAR. The project wasn't even called NEAR at the time; it was called NEAR.ai. Illia Polosukhin, the co-founder of NEAR, is one of the authors of the classic paper "Attention Is All You Need" published by Google and is the co-inventor of the modern Transformer architecture.
Kevin Follonier: When you invested in them in 2018, did you already realize that Transformer would become the absolute cornerstone of today's GPT and the entire generative AI wave?
Haseeb Qureshi: Of course not! If anyone today claims that they foresaw all this in 2018, they are simply making up history. In 2018, that paper was just an exploratory academic attempt on cutting-edge deep architecture published within Google Research. At that time, GPT-2 had not even been born, and no one on the market knew that large language models would detonate the world in this way in the future.
Haseeb Qureshi: The reason why we decided to re-focus is another story entirely. Illia and his co-founder Alex are both top-notch low-level system engineers. Alex previously worked in the distributed memory database MemSQL (now SingleStore). The two originally got together to work on AI projects. The reason why they switched to blockchain was because they studied the code of Ethereum and were completely stunned by the underlying implementation of Ethereum.
Haseeb Qureshi: Their intuitive feeling at the time was: "The code of Ethereum is too amateurish and terrible!" These guys have obviously never really built an industrial-scale distributed system. The current bottom layer of the blockchain is simply a joke. The two of us can definitely make a system that is ten or a hundred times better than this! 』
Haseeb Qureshi: So my partner at the time, Ivan, and I sat down with them and had a meeting. Illia drew a sketch for us on the spot on a piece of white paper, outlining the framework of the blockchain protocol they envisioned. At that time, there was an extremely popular concept in the industry, "Let blockchain nodes easily run and complete verification on smartphones" because everyone was anxious about decentralization and felt that if the node could not run on the mobile phone, it would not be decentralized enough. This requirement seems comical today, but at the time it was a huge narrative.
Haseeb Qureshi: Ivan and I stared at the plan he drew on paper and said frankly to them: "The idea of the plan is very novel, but from the perspective of engineering logic, it is full of holes. It does not work at all. The concurrency mechanism cannot be scaled at all. Some places are logically contradictory. We think this is by no means a good agreement." ’ So in that initial meeting, we rejected them face to face.
Haseeb Qureshi: It turned out that just one day later, Ivan and I received a long email from Illia and Alex early the next morning. The email read: "Thank you very much for your in-depth feedback yesterday. We have redesigned a new protocol solution, please check it out." 』They reconstructed and solved all the engineering flaws we picked out in just 24 hours!
Haseeb Qureshi: To be fair, the second version of the plan they came up with is still not perfect, but I immediately realized: the engineering ability of these two guys is simply unfathomable! In 2018, most of the people working on public blockchains in the industry were academics in the ivory tower, which were called "Professor Coins" at the time. They were all a group of theoretical scholars deriving formulas; and Illia and Alex were extremely rare hard-core Eastern European engineers, the kind of ruthless characters who could stay up late and push the code to the extreme in order to solve the underlying concurrency problems! If such a pair of top-level engineering teams are allowed to reconstruct the public chain, how much energy will be unleashed? It was this pure belief that led us to write our seed round check.
Kevin Follonier: What was the approximate valuation for that seed round? You just mentioned that the Solana seed round was 4 cents, what was NEAR at that time?
Haseeb Qureshi: My memory of the specific valuation figure is a bit fuzzy, but it must be far less than US$50 million, probably in the order of US$20-30 million.
Kevin Follonier: After eight years of evolution, what has NEAR become today? If you explain it to the public in the most straightforward language, what is its core form?
Haseeb Qureshi: After years of precipitation, NEAR has grown around a huge ecological galaxy, but its core pillar has evolved into a "full-chain cross-chain transaction and settlement center." It allows global users to seamlessly call liquidity on all mainstream blockchains and perform complex financial transactions through a unified account and address, under the premise of end-to-end privacy protection.
Kevin Follonier: As an early-stage venture investor, funds usually have an investment exit cycle. Has Dragonfly still held NEAR over the past eight years? Why choose not to sell for a long time?
Haseeb Qureshi: Yes, we are still heavily invested in NEAR. Just look at the historical price curve of NEAR to know how many thrilling ups and downs we have experienced during our holding period. But the reason why we firmly hold it is because we have unquestionable confidence in the team's background and technical reserves.
Haseeb Qureshi: In the underlying framework of venture capital, there are usually only two clear scenarios for selling assets: the first is that you have lost trust in the direction of the team and find that they have gone astray; the second is that the grand vision originally promised by the team has been 100% fully realized. They claimed to do
Haseeb Qureshi: But for NEAR, not only do we maintain absolute confidence in its team, but more importantly: their journey is far from over (The job is not done)! The early NEAR was once regarded as an Ethereum killer, trying to use sharding technology to land before Ethereum 2.0. At that time, Ethereum vowed to launch sharding in 2018, but it was postponed until after 2020. Even today, Ethereum has not been able to truly implement state sharding. However, NEAR did forcefully implement the sharding public chain back then.
Haseeb Qureshi: But even more valuable is the extremely low mental baggage displayed by the team. Many mediocre teams are extremely stubborn. For example, they originally set up an NFT project, but then the market collapsed, and they stubbornly sank with the sinking ship. The top teams have an extremely keen sense of reality. When they find that the market center of gravity has shifted, they quickly follow the pull of reality and turn to full-chain abstraction and distributed AI infrastructure. Almost all underlying protocols that have achieved great success in the history of encryption have their ultimate victory coming from this strategic evolution in the early stages.
The Rebellion’s Determination: Leading Ethena’s Investment Amid Terra Collapse and Industry-Wide Trauma
Kevin Follonier: Another masterpiece that you identified and invested in early on was Ethena. If I remember correctly, you just mentioned investing around 2022?
Haseeb Qureshi: To be precise, we are leading the seed round that was officially finalized in 2023. Because the epic Terra/Luna crash happened in May 2022, and Ethena’s financing unfolded shortly after the crash.
Kevin Follonier: After the Luna storm, the entire industry was mired in panic about post-traumatic stress disorder (PTSD), and the entire Internet became very enthusiastic about algorithmic stablecoins. In that extremely suffocating ruins of the industry, why do you dare to go against all the odds and lead the investment in Ethena?
Haseeb Qureshi: The general consensus in the entire venture capital circle at that time was that stablecoins were already a completely polluted forbidden area, and synthetic assets on the chain were absolutely not investable. Anyone who dared to raise money for stablecoins at this time was simply crazy. Therefore, when Guy Young was looking for seed round financing in the early days, he was rejected by countless institutions. Most investors didn't even bother to look at the mechanism. They were scared to stay away as soon as they heard "dollars on the chain".
Haseeb Qureshi: But the reason why we understand Guy is because we had previously invested in Ethena’s predecessor protocol, a project called Lemma Finance. Lemma was also trying to do basis arbitrage on the chain with spot plus perpetual short hedging, but they tried to complete the entire arbitrage hedging loop on pure decentralized derivatives DEX (such as dYdX and GMX).
Haseeb Qureshi: Lemma ultimately failed. The Achilles’ heel was that the capacity of the on-chain derivatives market at the time was too small, and the size of the on-chain open interest (Open Interest) simply could not support a multi-billion dollar stablecoin hedging position. We have been deeply involved in the field of on-chain perpetual contracts for seven years, and we know that on-chain derivatives will eventually rise, but at the time it was far from enough to quench our thirst.
Haseeb Qureshi: It was in that darkest moment that Guy found us. He analyzed the lessons of Lemma’s failure very profoundly: You can’t sit still and passively wait for the size of the perpetual contract on the chain to grow slowly. You must directly embrace the tens of billions of dollars of ready liquidity in centralized exchanges (CEX). The biggest drawback to embracing centralized exchanges is counterparty risk. What if the exchange runs away or is misappropriated?
Haseeb Qureshi: Guy pointed out keenly: The market structure has undergone irreversible qualitative changes after the crash! Due to the continuous thunderstorms of Terra, Three Arrows Capital (3AC) and FTX, centralized exchanges have encountered an unprecedented crisis of trust. In order to retain large market makers and institutional funds, major exchanges have begun to fully integrate third-party off-chain OTC custody and clearing solutions (Off-Exchange Settlement) for the first time, such as through Fireblocks, Anchorage and Copper ClearLoop to carry out OTC fund mapping.
Haseeb Qureshi: This means that large institutions mortgage their assets in the hands of compliant third-party custodians, and can directly place orders to establish short hedges without recharging funds into exchange accounts! A year earlier, it would have been impossible for the exchange to agree to such demanding custody requirements; but amidst the ruins of 2023, this historic door has been completely opened! Guy captured this once-in-a-lifetime structural gap, which directly untied all the knots in our hearts, so we led the seed round without hesitation.
Kevin Follonier: This seed round has been a while now. The investor’s token lock-up period should have passed long ago, right?
Haseeb Qureshi: Ethena’s token release was very candid. Although there was a normal lock-up period in the early days, the team actually completed a full early unlock for all early investors a few months ago. Today, all early investors, including us, have all their token shares in full circulation.
Regulating the Golden Channel Behind the Iron Curtain: The GENIUS Act and Ethena’s Desire for Global Yields
Kevin Follonier: Now that all shares have been unlocked, do you still not sell one share and continue to hold heavy positions? Why?
Haseeb Qureshi: Yes, we are still heavily invested. The reason is very pure: its growth ceiling is much higher than most people imagine.
Haseeb Qureshi: At the macro compliance level, the ongoing U.S. Stablecoin Act (GENIUS Act) has established a very critical iron law: all regulated legal currency-backed stablecoins (such as USDC) regulated in the United States are expressly prohibited from directly paying interest income to currency holders.
Haseeb Qureshi: Just imagine this world situation: when global inflation is high and benchmark interest rates remain relatively high, if ordinary people hold compliant stablecoins, the interest income they can obtain is strictly zero, which means an extremely huge opportunity cost of funds! Ethena's core product is essentially designed to provide stable dollar returns.
Haseeb Qureshi: With the increasing demand for U.S. dollar liquidity in countries in the Global South and non-U.S. regions, everyone is eagerly looking for a U.S. dollar savings vehicle that can bring real positive returns. As long as this global rigid desire for U.S. dollar returns exists, Ethena, with its natural exogenous basis return attributes, has every opportunity to continue to break through from today's billions of dollars to tens of billions or even hundreds of billions of dollars in the vast world of non-compliant fiat currency channels.
The Holy Grail of All-Asset Derivatives: Why the RFQ Model is the Only Solution for On-Chain RWA Perpetual Contracts
Kevin Follonier: In the derivatives track, you have recently participated in and led Variational. What is it about Lucas and his team that sets them apart? What underlying problem are they solving?
Haseeb Qureshi: Lucas is a rare genius. If I remember correctly, he was admitted to the university at the age of 12. He was extremely young but his intelligence was extremely amazing. When I sat down with him and had an in-depth conversation, you could immediately tell that the hardware configuration of this type of person was completely different.
Haseeb Qureshi: At the seed round, we wrote them a smaller check. They were actually working as a high-frequency market maker (Market Maker) at the time. We usually rarely invest in market makers because their business models tend not to scale like venture capital firms; but because of Lucas’ amazing talent, we made an exception and got involved. They later weathered the FTX crash and began exploring the next business model with long-term barriers.
Haseeb Qureshi: They have observed the rise of Hyperliquid and perpetual contract DEX on major chains, and are keenly aware that although the existing order book model performs well on crypto-native assets, it is not the optimal solution for future all-asset derivatives!
Haseeb Qureshi: In the crypto world, everyone is accustomed to setting up a separate set of order books every time a new exchange is opened, attracting market makers to repeatedly place orders in each independent buying and selling order. But in the traditional mature financial system, this is not how it works at all! In the U.S. stock market, when most retail investors trade on Robinhood, Charles Schwab or Fidelity, they do not match orders on an independent exchange built by the broker. The bottom layer of the broker uses a point-to-point request for quote clearing mechanism (RFQ, Request for Quote).
Haseeb Qureshi: Retail investors initiate a transaction request on the interface, and the backend system will push the request directly to Wall Street's top quantitative market makers (such as Citadel Securities, etc.). The major market makers will bid to provide the best quotes within milliseconds, and then complete cash settlement. This RFQ model forms the backbone of the entire modern retail brokerage business.
Kevin Follonier: What is the greatest value in moving this RFQ mechanism to an on-chain derivatives protocol?
Haseeb Qureshi: The greatest value is that it solves the Achilles’ heel of real-world assets (RWA) and massive long-tail assets being unable to establish liquidity on the chain! In the crypto circle, more than 80% of the perpetual contract trading volume is divided up by three or four mainstream tokens: Bitcoin, Ethereum, and Solana. The rest are extremely long-tail small currencies, so you can barely maintain order book depth on several mainstream chains.
Haseeb Qureshi:但在传统宏观金融世界中,根本不存在所谓的『绝对主流大盘股』!纳斯达克、标普指数、WTI 原油、黄金、以及数以千计的全球上市公司股票,全球投资者想要交易的资产代码多达数万个!你想想看,你怎么可能在区块链上为全世界上万种股票和商品分别搭建数万个深厚的独立订单簿并让做市商随时在链上挂满流动性?这在工程和资本效率上是绝对不可能完成的任务!
Haseeb Qureshi:如果一个海外交易员想在链上杠杆交易某只韩国股票,或者交易稀有金属钼(Molybdenum)的期货合约,一进交易所发现买卖差价高达 5%、盘口只有几百美元深度,他肯定转头就走、重新退回传统券商软件。唯有利用链上 RFQ 架构,只要链下有做市商愿意对锚定价格提供点对点清算,用户就能以与传统芝商所(CME)或币安一模一样紧密的买卖差价完成无滑点交易。这就是打开链上全资产衍生品的终极钥匙!
Kevin Follonier:你觉得这个赛道的潜在市场规模(TAM)究竟有多大?
Haseeb Qureshi:它的潜在空间是整个全球零售交易市场!我们可以把海外的差价合约(CFD)市场作为最直接的参照物。全球除了美国本土之外的大部分地区,比如在香港或者东欧,散户如果想杠杆交易美股或者全球外汇,绝大多数都是通过 CFD 差价合约平台进行的,因为他们无法直接触碰境外股票底层资产。
Haseeb Qureshi:全球 CFD 是一个体量极其惊人的万亿美元级庞大市场,而目前整个加密链上永续合约在其中占据的份额,连个位数的百分比都不到!更重要的是,传统离岸 CFD 券商往往充斥着严重的灰色操作,很多平台在塞舌尔等离岸岛屿注册,暗中执行 B-Book 模式,也就是平台直接作为对手盘和散户对赌,甚至暗中操纵插针吃掉散户本金。
Haseeb Qureshi:而在去中心化架构下,所有的清算逻辑、抵押品规则和智能合约代码全部公开透明地运行在链上。哪怕普通散户看不懂代码,在这个 AI 代理时代,你随便唤醒一个 AI Agent,它能在几秒钟内帮你完整审计智能合约的底层逻辑与历史交易流水。这种抗审查、无信任中介、只要一部手机和网络就能畅通无阻交易全球任何资产的基础设施,将彻底吞噬传统的离岸金融衍生品市场。
被 Hyperliquid 拒之门外的自省:当顶级团队拥有造血能力与自留 100% 价值
Kevin Follonier:聊到去中心化永续合约,我们不得不谈谈 Hyperliquid。这个项目从来没有进行过任何形式的种子轮风投融资,直接将代币全额分发给了早期社区与真实交易员。作为一家以寻找早期种子轮为使命的顶级风投,你们当时是怎么观察并尝试参与这个案子的?
Haseeb Qureshi:我们必须极其坦率地承认:我们不仅深度关注了,而且我们当时极度渴望能够投资他们!在 Hyperliquid 正式爆发并进行代币生成事件(TGE)之前,我和业内许多顶尖基金的合伙人都曾主动飞过去登门拜访创始人 Jeff,并为他开出了非常激进的投资要约。
Haseeb Qureshi:如果我没记错的话,Dragonfly 当时直接开出了高达 6 亿美元估值的投资报价,希望能够获得一部分战略份额。当时 dYdX 依然是全网体量最大的永续合约 DEX,但 Hyperliquid 的增长斜率已经极其迅猛,正在势如破竹地逼近王座。
Haseeb Qureshi:Jeff 面对我们开出的这份报价,认真权衡推演了一番,然后非常温和、得体但极其坚决地对我们说了声:『不,我们不打算接受外部资本。我们决定继续通过自有资金来支持开发,把网络全部的原生价值完整留给自己和社区。』据我所知,在他们临近 TGE 的中后期,甚至有同行开出了超过 10 亿美元估值的纯 VC 支票,同样被他毫不犹豫地一口回绝了。
Kevin Follonier:当你带着 6 亿美元的诚意扑了个空被当面拒绝时,你当时的心态是什么?这难道不会激起你更强烈的投资执念吗?
Haseeb Qureshi:当然会!当一个创始人指着自己的项目告诉你:『我不打算把风险转嫁给你们,我坚信用我自己的钱、以我的全部身家去押注我自己』的时候,任何一个真正懂行的投资人内心的反应一定是:『妈的,这家伙太有种了,我更想投他了!』
Haseeb Qureshi:在创业投资中,这才是顶级创始人最稀缺的优秀品质,全心全意押注自己。最让投资人感到心寒的糟糕信号,往往是那些在融资轮次中迫不及待地想要套现老股(Secondary)的创始人。如果一个创始人刚融到钱就迫不及待地套现几百万美元落袋为安,这释放出的潜台词极其负面:这意味着作为最清楚公司真实底牌的人,他其实对公司的未来毫无信心,他想尽快把风险转移给接盘的投资机构。
Haseeb Qureshi:而 Jeff 展现出的是完全相反的格局。他和他的核心量化团队从最早期起,就自掏腰包投入了数百万乃至数千万元的真金白银来支撑服务器和做市开发。而且从产品上线的第一天起,依靠他们极其精密的量化做市策略和爆发式增长的真实手续费流水,整个协议就具备了强大的正向现金流造血能力。
Haseeb Qureshi:当一个团队在资金上实现自给自足、现金流充沛、且产品日交易量正在迅速登顶全网第一时,他们到底有什么理由去凭空稀释宝贵的股权和代币给 VC 呢?Jeff 选择将 100% 的代币原生价值回馈给生态的真实交易者和流动性提供者,这不仅给传统风投行业好好上了一课,更树立了加密原生超级物种靠产品力和极致工程单兵突围的全新范式。
顶级投资人的退出哲学:在指数级赢家身上克服过早止盈的诱惑
Kevin Follonier:我之前在这个播客里采访过一位掌管数百亿美元的大佬,他曾说过一句非常经典的名言:『巨额的财富往往是在卖出的那一刻真正被锁定的。』面对像 Hyperliquid 或者你们捕获的其他顶级资产,在什么时候应该果断止盈?你们的退出哲学到底是什么?
Haseeb Qureshi:这很大程度上取决于资产的底层属性和生命周期。就像我前面提到的框架:对于平庸或者逻辑破产的项目,在失去信心或者预设论点被证伪时果断离场;而对于真正卓越的赢家,如果你过早地为了微薄的短期收益而卖出,往往会酿成整个基金生涯中最沉痛的战略失误。
Haseeb Qureshi:作为受托管理的风险投资基金,我们通常拥有长达大约十年的基金存续期。这意味着出资人(LP)赋予了我们充分的战略耐心与自由度:只要我们坚信这个赛道和这个项目在未来依然拥有远高于资本成本的超额内部收益率(IRR),我们就有充分的底气去死死拿住核心筹码。
Haseeb Qureshi:当然,作为合规专业基金,我们这些年也卖出过海量资产,为 LP 实现了极为丰厚的真实法币现金回流。如果一个风投只懂账面浮盈而从来不把真金白银分给 LP,你在行业里是绝对活不长久的。
Haseeb Qureshi:但 LP 最终追求的是全周期的超级阿尔法,而风险投资的底层规律是由极端的「幂次法则(Power Law)」支配的!以我们早期重仓的 Polymarket 为例,我们在其估值仅为 1.75 亿美元时就大举介入了。在过去几年里,市场上有无数个节点我们可以轻易地套现离场:拿到 10 倍回报卖一点,涨到 20 倍再卖一点,涨到 30 倍全部结清,账面上看这笔交易已经极其漂亮了。
Haseeb Qureshi:但正如最近媒体所披露的,Polymarket 最新的融资轮次估值正在朝着 20 亿美元乃至更高的超级巨头体量迈进!如果你在半山腰频繁地做聪明人、为了锁定浮盈而早早清空筹码,你实际上亲手封顶了一个时代级赢家所能为你带来的百倍、千倍指数级复利奇迹。对于那些真正构建起不可替代网络护城河的核心资产,最好的卖出时机往往是永远不卖。
押注 2027:隐私 AI 推理平台 Venice 与比特币的终极对冲保单 Zcash
Kevin Follonier:如果让你在当前所有的加密代币中挑出一个你目前信念最强、愿意坚定持有直至 2027 年甚至更久的核心标的,你会选择哪一个?
Haseeb Qureshi:如果非要选出一个当前在我心中占据第一优先级的资产,我的答案是 Venice。
Kevin Follonier:为什么是 Venice?很多大众可能对这个项目还不太熟悉,它到底在解决什么痛点?
Haseeb Qureshi:Venice 是一个基于去中心化加密架构构建的保护隐私的 AI 推理平台。它让全球任何用户都可以在完全无需许可、不受言论审查、且绝对不记录任何后端日志的前提下,直接调用全球最前沿的开源 AI 基础模型。你可以把它简单理解为一个完全匿名的去中心化版 ChatGPT。
Haseeb Qureshi:他们发行了自己的代币,拥有极其清晰扎实的商业模式:向全球 C 端用户收取高阶订阅费,并向开发者提供高性能的无日志推理 API 接口。这个业务正在以惊人的速度野蛮生长,月环比增长率高达 10% 左右,目前的年化营收运行率(ARR Run Rate)已经远远跨过了 1 亿美元大关!而在我们去年 12 月刚接触他们时,这个数字还仅仅只有 1500 万美元左右。
Haseeb Qureshi:虽然整个 AI 算力推理市场都在爆发,但 Venice 真正击中的是未来人类社会的生死痛点,AI 时代的个人隐私与公民自由。随着大模型深入每一个人的日常生活,它在本质上已经成为了我们人类大脑和意识的数字化延伸。但在现代法律体系面前,这种直觉是完全错位的。
Haseeb Qureshi:在美国司法体系中,存在一个著名的『第三方原则(Third-Party Doctrine)』:一旦你自愿将某项信息分享给第三方公司,法律就认定你已经主动放弃了宪法第四修正案赋予你的隐私期待权!这意味着只要你卷入诉讼,对方律师可以轻而易举调取你和 ChatGPT 的全部历史对话;执法机关甚至根本不需要向法官申请搜查令,只需开具一张普通传票,就能无死角审阅你在大模型里留下的所有商业机密、医疗咨询与潜意识思考。
Haseeb Qureshi:而放眼全球,中国在 AI 领域拥有极其严密的审查测试体系,任何模型上线前必须通过合规测试,大量历史和政治事实被严格过滤;在西方,Anthropic 最近也在 Claude 中加入了敏感词拦截规则,一旦检测到所谓的情绪化攻击词汇就会强行切断对话。未来这个禁令名单还会被塞进多少东西?
Haseeb Qureshi:面对这种密不透风的中心化监控与思想审查,每一个拥有独立意志的人类内心深处都会自然涌现出一股最原始的愤怒与反叛:『滚你的!由我来决定我对 AI 说什么,而不是由你们这些寡头来规训我怎么思考!』这种普罗米修斯式的自由火种是深深植根于人性的。Venice 扮演的正是 AI 时代的去中心化无日志 VPN,它是未来人类守护思想自由的不可退让的根本防线。
Kevin Follonier:除了 Venice,你们还有一个持有长达近十年但极具争议的老牌资产:Zcash。外界很多人对你们为什么重仓 Zcash 感到大惑不解,你的底层投资逻辑到底是什么?
Haseeb Qureshi:是的,Dragonfly 确实重仓持有 Zcash。要理解这个资产在当下的颠覆性价值,可以引入两个全新的核心维度。第一个维度是 Naval Ravikant 曾提出过的经典论断:Zcash 在本质上是你为你持有的比特币购买的一份终极保险单!
Haseeb Qureshi:从极简主义的角度来看,Zcash 没有复杂的智能合约生态,它本质上就是『具备不可穿透隐私特性的纯净版比特币』。如果你认同非主权硬通货的历史地位,你就必须清醒地意识到:非主权资产最核心的价值基石在于它绝不能被任何人轻易没收或冻结。但今天的比特币正在走向高度的合规化与账本全透明化,监管机构正在对所谓的『被污染 UTXO』实施全面拉黑与拦截,比特币的抗审查可替代性正在遭受不可逆的侵蚀。
Haseeb Qureshi:而第二个维度,则是最近引发全网震动的密码学生存危机,AI 正在以不可思议的速度解构基础数学!就在不久前,OpenAI 披露其尚未公开的内部前沿推理模型,在短短 4 个小时内接连攻克了 700 多项世界顶尖数学难题。而在两年前,大模型连高中的基础代数题都经常做错,今天它已经开始尝试求解纳维-斯托克斯方程(Navier-Stokes)这一千禧年难题!全球数学界公认这是人类历史上单日取得数学突破最庞大的一天。
Haseeb Qureshi:最新流出的动向是:OpenAI 正式将这套强大的算力激光直接瞄准了现代密码学的底层根基!全球密码学界正在陷入巨大的恐慌之中。一旦经典的椭圆曲线公钥加密在数学上被 AI 撕开一道裂口,整个全球数字金融将面临灾难性的『数学浩劫(Mathocalypse)』!由于比特币的社区治理共识极其笨重和保守,一旦遭遇底层数学崩溃,它很难在短时间内迅速完成全网硬分叉升级。
Haseeb Qureshi:相比之下,Zcash 社区早已在后量子密码学和高阶零知识证明防御上建立起了极度完备的应对预案。再加上比特币在 2140 年前乃至未来几次减半后区块奖励骤降所带来的安全预算危机,任何一个打算持有比特币十年以上的理性投资者,都应该按照自己对尾部风险的担忧程度,买入相应比例的 Zcash 作为对冲保单。
绝境中的超级多头:什么情况会让 Haseeb 转向看空?以及硅谷折腾文化的不可替代性
Kevin Follonier:在加密行业每当行情低迷、大家垂头丧气时,你总是那个极其坚定的超级多头。从一个反思的角度来看,究竟发生什么样的情况,才会真正让你从根本上转向看空加密行业、甚至选择离开?
Haseeb Qureshi:有两件事情会真正让我彻底转向看空。第一,是对加密核心价值观的彻底背叛。如果有一天加密技术彻底放弃了它作为『抵抗中心化剥削、为个体提供逃离传统体制的避风港』的立身之本,全面向华尔街投降,退化成一个纯粹为了迎合监管的无聊金融玩具,那我对它的信仰将会坍塌。
Kevin Follonier:但现在贝莱德(BlackRock)推出了现货 ETF,传统机构全面涌入,很多人认为这本身就已经背离了赛博朋克的精神初衷,你认同这种观点吗?
Haseeb Qureshi:我完全不认同这种肤浅的解读。贝莱德发行比特币 ETF,难道会妨碍你在底层链上自主转账、或者妨碍你在 Zcash 里保护自己的隐私吗?就像一个巨大的公共游泳池,有人在泳池的那一头游泳,并不妨碍你在这一头畅游。
Haseeb Qureshi:关键的本质在于:从来不是赛博朋克被迫运行在传统机构的铁轨上,而是传统机构最终不得不屈尊运行在去中心化密码学的底层轨道之上!除非有一天所有的出块矿工和验证者全部被贝莱德彻底收买并只执行审查指令,否则这种机构化只会让密码学底层更加坚不可摧。
Haseeb Qureshi:而第二件会让我转向看空的事情,是加密行业彻底丧失了自我进化的创造力。这种创造力的消亡往往不是因为人才变笨了,而是因为文化遭到了破坏,当人们开始对尝试疯狂的新事物冷嘲热讽、对失败展开道德审判时,创新就彻底死了。
Kevin Follonier:这确实是欧美与亚洲传统文化的一大痛点。比如在欧洲很多国家,哪怕像瑞士这样自诩高度创新的体制,其实社会的风险偏好极度保守,根本没有人愿意承担失败的风险,创业失败往往会被社会视为一辈子的职业污点;在新加坡同样如此,所以这些地方几乎诞生不了世界级的科技巨头。
Haseeb Qureshi:这正是硅谷文化最不可替代、最神圣的地方。每个国家都在宣称要复制一个硅谷,但当他们得知硅谷成功的真正代价,你必须允许疯狂的人去尝试荒唐愚蠢的新想法,允许他们把大笔资金挥霍一空并彻底爆炸,而且当他们失败归来时,整个社会不仅绝不嘲笑审判他们,甚至将他们视为探索未知边界的英雄,一旦听到这个代价,所有的传统体制全部吓得连连后退。
Haseeb Qureshi:以红杉资本为例,投资 FTX 毫无疑问是红杉整个五十年历史上最惨痛、最丢脸的道德与财务污点之一。在 FTX 崩盘之后,大部分卷入其中的传统资本发誓再也不碰加密货币;但红杉今天依然在坚定不移地投资 Web3 和加密底层!他们本可以彻底退回只做 AI 和生物医药的舒适区,但他们依然选择回到赛场。
Haseeb Qureshi:这就是最纯粹的硅谷灵魂:被迎面重重打了一拳、鼻青脸肿地倒在地上,第二天清晨拍拍身上的尘土,若无其事地站起来继续战斗!正是这种生生不息、鼓励离经叛道的折腾韧性,孕育了个人电脑、互联网、开源运动、加密货币以及今天的 AI。只要这种神圣的文化还在,任何短期的熊市低谷与监管风暴,都注定无法阻挡技术奔涌向前的滔天巨浪。
风险提示
根据中国人民银行等八部门最新发布的银发〔2026〕42号文件,本平台内容仅用于区块链行业技术与信息分享,不构成任何投资建议,不推广、不背书任何经营与投资行为。请读者严格遵守所在地区法律法规,谨防投资风险与诈骗。目前存在大量诈骗类仿冒账号,敬请警惕甄别。本平台所有内容未经许可,禁止转载、复制及商用。
北京蓝色海马文化传播有限公司 版权所有 京ICP备2023026470号




