Written & Hosted by Mr. Z (@168MrZ)
Cards have already captured Agent payments; Crypto's real value is shifting toward privacy, financial autonomy, and Agentic UX; Hyperliquid is expanding the market rather than simply cannibalizing CEXs, and Robinhood Chain still has to prove RWA and the developer ecosystem
Monday, October 12, 2026, 2:00 PM. 168X once again invited Dragonfly Managing Partner Haseeb Qureshi (@hosseeb). It has been more than half a year since we last met in Hong Kong. The Crypto market has gone through an obvious drawdown and has begun to recover its temperature again. Crypto has been a bit boring this year, but he thinks the Bull Market is coming back.
Haseeb, just back from TOKEN2049 Singapore, is no longer seeing the Crypto of the previous cycle driven by Parties, Memes, and pure Retail Speculation. TOKEN2049 still drew about 25,000 people this year, but one detail that says more is this: in past Token Weeks, the Marina Bay Sands Marquee was almost booked out every day by Crypto projects, yet this year it was just operating normally for the entire week, with no Crypto project booking it out at all.
The truly important change happened in AI. At the beginning of the year, Haseeb also agreed with a very popular Thesis: when AI Agents truly become Autonomous, they will naturally need Stablecoins and Crypto Rails. But now that Agents have actually arrived, the answer is completely different: Agents did not wait for Crypto; they used Cards first.
In this conversation, we went from AI Agents, Privacy, and Agentic Finance all the way to Robinhood Chain, Hyperliquid, HIP-3, Equity Perps, Dragonfly's next stage, and why Haseeb believes that in the future there may not even be a category called "Crypto VC."
1. TOKEN2049's most honest indicator: the Marquee was not booked out for the first time
Mr. Z: You just got back from TOKEN2049 Singapore. What was the biggest difference between the event this time and what you had originally expected?
Haseeb: TOKEN2049 should now be the world's largest Retail Crypto Conference, with about 25,000 people this year. But if you ask me what best reflects the current state of Crypto, I would instead look at the Marquee. In every past Token Week, the Marquee at Marina Bay Sands was almost booked out every night by some Crypto project. But this is the first time I have seen the Marquee just be a normal Nightclub for an entire week, with no Crypto project booking it out.
Haseeb: That says a lot about how the industry has changed. Retail Activity, Volatility, and project hype are all still there, but the Energy of the crazy Speculation from past Cycles is clearly much lower. On the other hand, one very important event at TOKEN2049 this year was a Closed-door Institutional Event, with many Hedge Funds, Asset Managers, Founders, and Regulators flying to Singapore specifically to attend.
Haseeb: Conferences still have Alpha. By constantly talking to people, you still learn what capital has been looking at recently and which projects are gaining Attention. But the industry as a whole is indeed getting older, more mature, and more Institutional. In the past, many people came to Party; now more and more people are really coming to find Business, Deals, and Flow.
2. AI Agents did not wait for Crypto at all: Cards won payments first
Mr. Z: In your talk at TOKEN2049, you publicly revised one view. At the beginning of the year, many people said AI Agents would naturally need Crypto, but now you say they do not need it. What made you change your mind?
Haseeb: At the beginning of the year, many people, including myself, thought Crypto would become necessary Infrastructure for AI. The logic was intuitive: an AI Agent has no driver's license and no Social Security Number, so how does it KYC, how does it open a bank account, and how does it use traditional Payment Rails? So we naturally deduced that Stablecoins would become the main payment method for Agents.
Haseeb: The problem is that at the beginning of the year, Agents were still quite theoretical. Now they are no longer theoretical. Both Enterprise and Consumer Adoption are starting to grow rapidly, so we can finally directly observe what they are actually using. The answer is simple: they are mainly using Cards.
Haseeb: Visa, Mastercard, and companies like Rain, Ramp, and Stripe built on top of the Card Network quickly built what Agentic Payment needed. Spending Limits, Risk Limits, Permissions, all of these can be put directly into existing Card Rails. We originally thought traditional finance would not adapt so quickly, but they solved this problem faster than we imagined. So this should be directly acknowledged: we got it wrong.
3. Crypto does not need to become the default payment layer for Agents: it is only called when it is truly better
Mr. Z: If Cards can already solve a large amount of Agent Payment, what truly Defensible advantage does Crypto still have left?
Haseeb: Agents will use Crypto for roughly the same reason humans do, which is when it is the best tool for completing the task. If I want to make an international remittance and Stablecoins are cheaper and faster, then my Agent will naturally discover that it should use Crypto. If some On-chain Market offers a function that traditional finance does not, it will use that too.
Haseeb: The mistake was that we used to think: "Because it is Software, it must prefer Crypto." Actually, it will not. Agents have no ideological loyalty to Crypto; they only Optimize. Sometimes the answer is Card, sometimes API, sometimes Stablecoin, and sometimes On-chain Protocol.
Haseeb: I think this Thesis is actually healthier. Crypto does not need to win all payments; it only needs to win where its architecture truly has an advantage. Ultimately the market will force Crypto to answer a more basic question: exactly which financial behaviors are simply better with Blockchain than with traditional Rails?
4. The real AI & Crypto intersection: Agents will become your "Crypto Guy"
Mr. Z: If Payments is not the biggest story, then how exactly will AI change Crypto?
Haseeb: The biggest change will happen in Interface. Today you still have to Bridge yourself, watch Gas, find Yield, compare Aave and Morpho, Rebalance, and find the lowest Trading Fee. In the future, many of these things will disappear, and you will only need to tell the Agent: "What I want to accomplish."Haseeb: Your Agent will become your Crypto Guy. It helps you find the best Yield, the lowest fees, the best buying routes, and it will also do Due Diligence for you. Every time it interacts with a Smart Contract, it can check the contract, check the team, look at LinkedIn, search X, and determine whether there is a Honeypot, Ponzi, or Rug Risk, and it will never skip checks out of laziness, nor will it FOMO.
Haseeb: Today's Agents of course still make mistakes, but today is also the worst they will ever be in the future. With every Model Release, they will only get better. Just like Self-driving Cars, 5 years ago they might not have been as safe as humans, but today is already completely different. So the truly interesting conclusion may not be that AI makes Crypto more suitable for machines, but that AI finally makes Crypto safe enough and simple enough for ordinary people to actually use.
V. Privacy is the real Killer Use Case: the stronger AI becomes, the more individuals need Crypto
Mr. Z: Dragonfly recently invested in Venice AI (@AskVenice) and Sapiom (@sapiom). Where do you think the core intersection of AI and Crypto really is now?
Haseeb: I think the most important thing is Privacy. AI itself is actually very dangerous for privacy, because every time you hand data to AI, it usually means handing data to a third party. Psychologically, we feel that our conversations with AI are very private, like an extension of our own brain, but from a technical and legal perspective, it does not necessarily belong to your own Private Space.
Haseeb: AI will also make large-scale Surveillance much easier. Data that used to be scattered in different places can now be instantly integrated, analyzed, and even De-anonymized by models. So as AI strengthens the capabilities of states and large institutions, there also needs to be another set of Technology to strengthen the capabilities of individuals. This is the meaning of Privacy-preserving Cryptography, Zcash, Monero, and Private AI Infrastructure.
Haseeb: Venice is a very good example. You can of course use traditional Login and Credit Card, but if you really want strong privacy, you can use a Crypto Wallet, use Crypto Payment, and combine that with Encrypted Inference. If you use a Credit Card, the platform will definitely know who you are. But if you use Crypto, plus true End-to-end Encryption, then the platform may not even know your name, location, or what Model you use.
Haseeb: So now I would say something very different from the beginning of the year: Crypto is not for your AI Agent to use, it is for you to use. Your Agent merely uses it on your behalf.
VI. What Agentic Finance will eliminate first is the "laziness tax" of banks and brokerages
Mr. Z: You have also recently been discussing the Agentic Bank Run. If everyone's Agents start automatically managing money, will that make the financial system more fragile?
Haseeb: What happens in the first stage may not be a Bank Run at all, but rather the destruction of many Business Models built on human laziness. Suppose you put a large amount of Cash in a JPMorgan Checking Account, with a very low interest rate, even though Money Market can clearly give you a much higher Yield. Why doesn't everyone move their money every day to the place with the highest return, and move it back when they need to pay? The answer is simple: because people are lazy.
Haseeb: Agents will not be lazy. It can work 24/7, check your funds once an hour, and automatically move Idle Cash to the optimal place. Cheap Deposits are one of the most core parts of the business model of many large banks, because consumers are too lazy to move their money, so banks can lend out very cheap funds and earn the Spread. Many Brokers are the same. A large part of the Economics of Schwab and Fidelity comes from clients leaving Cash in Sweep Accounts.
Haseeb: Agents will arbitrage away this "human laziness tax." What is most likely to happen is not that all banks are suddenly run, but that banks are forced to raise Deposit Rates to retain funds. As a result, Funding is no longer so Cheap, and the Margin and ROE of banks and Brokers will both be compressed.
VII. Amazon has already started blocking Agents: Agent-first and Local Agents will win
Mr. Z: But these Incumbents will not do nothing and just wait for Agents to take away their profits, right?
Haseeb: Of course not. The second path is to directly Block Agents. This has actually already begun to appear. If a platform's profits come largely from Advertising, Sponsored Placement, or User Friction, then an Agent that truly represents the user's interests is itself a threat. Because the Agent does not care at all what a Sponsored Result is, it will only look for the best result for the user.
Haseeb: Banks will have the same Incentive in the future. They will not say, "I do not want your Agent to move Deposits to a place with higher Yield." They will say "this is not safe," "there is Compliance Risk," "AI may Hallucinate." But the underlying conflict of interest is actually the same.
Haseeb: So in the long run I am especially bullish on two types of things. The first is Agent-first Companies, which are willing to accept Agentic Commerce from day one. The second is Local Agents. If the Agent you use essentially belongs to Meta or another third-party platform, the other side can Block that company; but if the AI truly runs on your own computer and operates for you through a Browser like a human, then it is very hard for them to Block you.
Haseeb: Ultimately, the biggest power shift is from "the platform decides what you can do" to "your Agent truly acts on your behalf in your interests."
VIII. 2026 prediction review: BTC at 150k is difficult, but Fintech Chains were indeed overestimated
Mr. Z: At the end of last year you made a long list of 2026 Predictions. Looking back now, which judgments were the most accurate, and which need revision?
Haseeb: Overall I think it is not bad, but of course there are also mistakes. At the time I expected BTC to reach 150,000 by the end of the year. From where it is now, getting to 150,000 by the end of the year would require nearly doubling, so the difficulty is already very high. But I think it is still possible to get back above 100,000.
Haseeb: On Fintech Chains, I think the direction is basically correct. Last year the market once believed that once Tempo, various Fintech Chains, and Robinhood Chain came out, they would use Distribution to suck away a large amount of Stablecoins and RWA, and even challenge Ethereum and Solana. But so far, the scale everyone imagined has not appeared.Haseeb: Robinhood is still very early, so I won't jump to conclusions. But the real surge in new RWA is still mainly happening on Ethereum and Solana. Distribution matters, but Distribution by itself does not automatically create a Blockchain Ecosystem.
9. Robinhood Is Not the Next Solana Yet: What's Really Missing Is RWA and Developers
Mr. Z: I want to push back on this. Robinhood Chain recently already has Agent Finance, Programmable Hooks, Stock Token Liquidity, Perps, and some new Market Structure. Do you still think it's far from Solana?
Haseeb: I think Robinhood Chain is still TBD (To Be Determined). There is indeed a lot of Meme Activity right now, but Robinhood was never meant to be a Meme Chain. What it really wants to tell is the RWA story. And right now real onchain RWA Activity is still very small, and its share of the entire RWA market is very limited.
Haseeb: Another issue is Developers. I currently don't see many truly independent developers continuously building new Applications and new Ideas on Robinhood. Right now it's more Meme Farming, Meme Cycle, and some Market Structure Experiments. Of course, Perps, DeFi Market Structure, and these things are worth watching, but putting absolute scale back into the comparison, Robinhood Chain is still much smaller than Hyperliquid, Arbitrum, BSC, Tron, Base, and Solana.
Haseeb: I'm not saying it definitely won't succeed. I just think the market previously directly inferred from "Robinhood has Distribution" that "Robinhood Chain will swallow everything," and to this day that has not been proven. What's really worth tracking is whether it can later produce sustained growth in RWA Activity, Independent Developers, and Native Applications that can form a Flywheel without relying on the Robinhood brand itself.
10. PerpDEX Will Not Be Left with Only Hyperliquid: Equity Perps Are the Structural Opportunity
Mr. Z: You previously predicted that PerpDEX would eventually Consolidate into a few large Venues. Now that Hyperliquid is so strong, does that judgment still hold?
Haseeb: If you really look at the numbers, it's actually not that far from what I originally imagined. In a 24-hour Snapshot today, Hyperliquid Reported Volume is about $4 billion, while Variational (@variational_io) and Lighter (@Lighter_xyz) are each about $1.5 billion. Of course, it's still not the 40 / 30 / 20 I originally predicted, but a structure of a few large Venues has already begun to emerge.
Haseeb: We are investors in Variational and Lighter, and we also hold some HYPE, so this needs to be disclosed. But I still believe that as the Liquidity Flywheel of these Venues becomes more mature, the market structure will continue to Consolidate.
Haseeb: Another thing I'm fairly confident I got right is Equity Perps. At the beginning of the year this market was basically close to 0, and now it has clearly taken off. I think one very important growth curve for DeFi Derivatives in the future is bringing stocks, and more Traditional Assets, into the same Perpetual Infrastructure.
11. HIP-3 Is Not as Permissionless as Imagined: TradeXYZ (@tradexyz) Looks More Like a Deep JV
Mr. Z: I personally really like HIP-3. Jeff is essentially telling everyone that you can come to Hyperliquid, open your own Casino, set your own rules, as long as you pay Rent. Why are you instead somewhat Skeptical of this Model?
Haseeb: Because if the HIP-3 Model were naturally this effective, then before TradeXYZ appeared, we should have already seen many successful cases. But that's not the case. TradeXYZ is indeed a huge success, but it is not some random team walking in through the Front Door and then automatically succeeding through HIP-3.
Haseeb: It already had a relatively deep History with the Hyperliquid Ecosystem, including previous collaboration in areas such as Spot Infrastructure. So this feels to me more like a Deep Partnership, even a Quasi-JV, rather than proof that a completely Neutral Permissionless Marketplace will spontaneously grow great companies on its own.
Haseeb: If you really want to Build a very large Market, you also run into a very practical problem: why should I build my entire company on someone else's land, pay a large amount of Rent over the long term, and at the same time accept the Platform Risk that the other side can change the Rules in the future? If you are the next Polymarket, you will very likely want to own your own Infrastructure.
Haseeb: Hyperliquid can give you very good Rails, but it will not automatically do Market Development for you. A brand-new category from 0 to 1 still requires a Founder, a complete Team, Capital, and Distribution. Protocol Primitive can lower execution difficulty, but it cannot replace Company Building.
12. Hyperliquid Is Not Eating CEX: It Has Actually Created a New Market
Mr. Z: The most common claim in the market is that Hyperliquid is taking Market Share from Binance, OKX, Bybit, and Bitget. Do you think this claim is not entirely correct?
Haseeb: If you really analyze the Addresses on Hyperliquid and compare them with CEX Usage, you'll find that the overlap between users on the two sides is much lower than people imagine.
Haseeb: Three or four years ago, there were still many Non-KYC CEXs globally, and the User Base of those Exchanges may be closer to today's Hyperliquid. But now all major CEXs are highly Regulated and basically all require KYC. There is a group of users who, because of Region, KYC, Regulation, or other reasons, never had good CEX Access to begin with. For them, Hyperliquid is not "moving over from Binance," but finally having a place where they can trade.Haseeb: If Hyperliquid were really eroding CEXs, then you should see Hyperliquid Volume rising while CEX Volume declines. But what you're seeing now is both sides growing. So I think right now it's closer to Market Expansion than Market Share Transfer.
13. HYPE's Next Growth Curve: Pre-IPO Has PMF, but Crypto Core Is Still the Base
Mr. Z: Traditional Market will eventually get closer and closer to 24/7 Trading. Once "round-the-clock trading" is no longer a Moat, where is Hyperliquid's next growth curve? Pre-IPO, Private Markets, HIP-3?
Haseeb: Pre-IPO Equity clearly has Demand, and it also has Product-market Fit. TradeXYZ also still has a lot of room to continue expanding. But if you ask me what Hyperliquid's next biggest Product Curve is, I don't have a very certain answer right now.
Haseeb: That said, I think people sometimes get too obsessed with "how Hyperliquid breaks out of Crypto," and instead overlook that Hypercore's own Crypto Business is already very strong. I'm relatively bullish on the Crypto Market over the next 12 months, so if Crypto continues to recover, I think HYPE itself, as well as the Exchange built around Crypto Trading, both still have a good environment.
Haseeb: The ideal outcome is not for Hyperliquid to leave Crypto, but for Crypto Core to continue Compounding, while Equity, Pre-IPO, and other Financial Assets expand the TAM layer by layer.
14. Dragonfly Is No Longer Just an "East-West Bridge": The Next Stage Depends on a Complete View of Global Crypto
Mr. Z: Dragonfly's earliest important positioning was Bo Feng, Mr. Feng Bo (@FengBo22588938), connecting China's first generation of internet capital with global Crypto. Today the China-US environment is completely different. Should Dragonfly still be understood as an East-West Bridge?
Haseeb: When I first entered Crypto, China was actually relatively Constructive toward the entire industry. In 2018 there was of course the ICO Crackdown, but Crypto wasn't completely hidden away. By 2020, when the market entered a phase of rapid growth, there was also a feeling that Crypto would ultimately become a globally accepted Technology.
Haseeb: After 2021, China banned Crypto and Mining, and many Founders who originally Built in China moved to Singapore, Hong Kong, and later Dubai and the US. So today I wouldn't say Dragonfly's job is just to connect China and the United States. Crypto itself is Global.
Haseeb: The Chinese Diaspora is still a very important part of this industry, it's just that they have already dispersed across different countries. Our LPs also include Asian capital, as well as US Institutions. I often say that if you really want to do Crypto Investing well, you have to be able to see the whole Elephant, not just feel one part of it like the blind men.
Haseeb: This is also why we have a presence in both Singapore and New York. We want to see Founders who are truly creating value in different parts of the world, rather than first deciding whether something is worth investing in based on where the company is registered. I have also long believed that China will eventually have a moment when it realizes again that Crypto has become too big to completely ignore. It is too important, and too much like a foundational Technology belonging to the 21st century.
15. Crypto VC Will "Disappear": What Remains in 2030 Are the People Who Truly Know How to Invest
Mr. Z: You said before that as Crypto Matures, "Crypto VC" may disappear in the future. By 2030, what kind of Fund will still qualify to exist?
Haseeb: I'm not saying all Crypto Investment Firms will disappear, but we won't call it a Crypto Fund anymore. Crypto has already entered the Water Supply. Stablecoin will become Fintech Backend, Neobanks will integrate Crypto, Consumer Products will have built-in Wallets, and more and more tech companies will quietly use Blockchain Rails in the background.
Haseeb: There used to be so-called "Internet Investors." When the Internet first emerged, that category of course made sense. But later, when all companies became Internet Companies, there was no longer so-called Internet Investing, only Investing. I think Crypto will eventually reach the same state.
Haseeb: This Transition will be very brutal for Funds whose only Edge is "I know how to buy Tokens." In 2020 and 2021, you might Raise a Fund, buy a few Tokens, Flip some Shitcoins, and you could call yourself a Crypto VC. It will get harder and harder going forward.
Haseeb: The real Durable Edge will become: Do you actually understand this world? Do you have real Relationships? Can you help a Founder solve problems? Beyond "you understand Crypto," what else can you bring? If you have nothing else, then you won't Survive this Transition. If you do, then I actually think the opportunities in the future are very large.
Mr. Z: So now that you've come to Asia, who do you most want to meet: Researchers, Investors, or other people?
Haseeb: Always Founders. I want to meet people who are Building new things and have new Ideas. We have always believed that Great Ideas can emerge from anywhere in the world.
Mr. Z: At the end of this episode, I think it can be condensed into 4 judgments. First, AI Agents have not treated Crypto as the default Payment Rail as the market imagined at the beginning of the year; Cards have more quickly captured a large amount of Consumer Payments. Second, the real opportunity in AI & Crypto may actually be even bigger; Agents will rewrite Crypto UX, and Privacy will become important Infrastructure for protecting individual autonomy in the AI era. Third, Hyperliquid currently looks more like it is creating a new Market rather than simply taking Volume from CEXs, and at the same time HIP-3 itself will not automatically create a great company. Fourth, Crypto is transforming from an independent industry into the underlying Infrastructure of the entire financial and technological world, and in the end even the category of "Crypto VC" may lose its meaning. Thanks to Haseeb for coming to 168X again!