Conversation with Dragonfly Partner Haseeb: Agents Don't Need Crypto, Humans Do!

Written by Mr. Z (@168MrZ)

Cards have already taken over Agent payments, while the true value of Crypto shifts towards privacy, financial autonomy, and Agentic UX; Hyperliquid is expanding its market rather than merely eating into CEX, and Robinhood Chain still needs to prove RWA and its developer ecosystem.

On October 12, 2026, Monday at 2 PM, 168X once again invited Dragonfly Managing Partner Haseeb Qureshi (@hosseeb). It has been over six months since we last met in Hong Kong, and the Crypto market has experienced a notable pullback, but it is starting to regain momentum. Crypto has been a bit dull this year, but he feels that the Bull Market is returning.

Just back from TOKEN2049 Singapore, Haseeb sees that the Crypto landscape is no longer driven by parties, memes, and pure retail speculation as it was in the last cycle. TOKEN2049 still attracted about 25,000 attendees this year, but a more telling detail is that the Marina Bay Sands Marquee, which used to be booked out by Crypto projects almost every night during Token Week, was just operating normally this week without any Crypto project taking over.

The truly significant change is happening in AI. Earlier this year, Haseeb agreed with a very popular thesis: when AI Agents truly become autonomous, they will inherently need Stablecoins and Crypto Rails. But now that Agents have actually arrived, the answer is completely different: Agents did not wait for Crypto; they first used Cards.

In this conversation, we discussed AI Agents, Privacy, Agentic Finance, Robinhood Chain, Hyperliquid, HIP-3, Equity Perps, the next phase for Dragonfly, and why Haseeb believes that in the future, there may not even be a category called "Crypto VC" anymore.

1. TOKEN2049's Most Honest Indicator: Marquee Was Not Booked for the First Time

Mr. Z: You just returned from TOKEN2049 Singapore. What was the biggest difference from what you originally expected?

Haseeb: TOKEN2049 should now be the largest Retail Crypto Conference in the world, with about 25,000 attendees this year. But if you ask me what best reflects the current state of Crypto, I would look at the Marquee. In past years during Token Week, the Marina Bay Sands Marquee was almost booked every night by some Crypto project. But this is the first time I’ve seen the Marquee operate as a normal nightclub for an entire week without any Crypto project taking over.

Haseeb: This clearly indicates that the industry has changed. Retail Activity, Volatility, and project hotspots are still present, but the crazy speculation energy from previous cycles is noticeably less. Conversely, an important event at this year's TOKEN2049 was a closed-door institutional event, where many hedge funds, asset managers, founders, and regulators flew to Singapore to participate.

Haseeb: Conferences still have Alpha. By chatting with people, you can learn what funds are currently looking at and which projects are gaining attention. But the entire industry is indeed aging, maturing, and becoming more institutional. Many people used to come for parties, but now more are genuinely looking for business, deals, and flow.

2. AI Agents Didn't Wait for Crypto: Cards Won the Payment Battle First

Mr. Z: In your speech at TOKEN2049, you publicly corrected a viewpoint. Earlier this year, many said AI Agents would inherently need Crypto, but now you say they don’t. What made you change your mind?

Haseeb: Earlier this year, many, including myself, believed that Crypto would become essential infrastructure for AI. The logic was straightforward: AI Agents don’t have licenses or Social Security Numbers, so how can they KYC, open bank accounts, or use traditional Payment Rails? Thus, we naturally deduced that Stablecoins would become the primary payment method for Agents.

Haseeb: The problem is that, at the beginning of the year, Agents were still very theoretical. Now they are no longer theoretical; both enterprise and consumer adoption are rapidly increasing, so we can finally observe what they are actually using. The answer is simple: they are primarily using Cards.

Haseeb: Companies like Visa, Mastercard, and those built on Card Networks like Rain, Ramp, and Stripe have quickly created what is needed for Agentic Payments. Spending Limits, Risk Limits, Permissions—these can all be directly integrated into existing Card Rails. We originally thought traditional finance wouldn’t adapt this quickly, but they solved this issue faster than we imagined. So this point should be acknowledged: we judged incorrectly.

3. Crypto Doesn't Need to Become the Default Payment Layer for Agents: It’s Only Called Upon When Truly Superior

Mr. Z: If Cards can already solve a large number of Agent Payments, what real defensible advantages does Crypto have left?

Haseeb: Agents will use Crypto for reasons similar to humans, which is when it is the best tool for completing a task. If I want to make an international remittance and Stablecoin is cheaper and faster, then my Agent will naturally realize it should use Crypto. If a certain on-chain market offers functionalities that traditional finance does not, it will also use it.

Haseeb: The mistake was in thinking, "Because it is Software, it must prefer Crypto." It actually doesn’t. Agents have no ideological loyalty to Crypto; they only optimize. Sometimes the answer is Cards, sometimes it’s APIs, sometimes it’s Stablecoins, and sometimes it’s on-chain protocols.

Haseeb: I think this thesis is actually healthier. Crypto doesn’t need to win all payments; it just needs to win where its architecture truly has advantages. Ultimately, the market will force Crypto to answer a more fundamental question: which financial behaviors are better with Blockchain than with traditional Rails?

4. The Real Intersection of AI & Crypto: Agents Will Become Your "Crypto Guy"

Mr. Z: If Payments aren’t the biggest story, how will AI change Crypto?

Haseeb: The biggest change will occur in the Interface. Today, you still have to bridge, check gas, find yield, compare Aave and Morpho, rebalance, and look for the lowest trading fees. In the future, many of these tasks will disappear; you will just need to tell the Agent, "I want to accomplish this."

Haseeb: Your Agent will become your Crypto Guy. It will help you find the best yield, the lowest fees, and the best buying paths, and it will also conduct due diligence for you. Every time you interact with a Smart Contract, it can check the contract, the team, look at LinkedIn, search X, assess whether there’s a honeypot, Ponzi, or rug risk, and it will never skip checks out of laziness or FOMO.

Haseeb: Today’s Agents will certainly still make mistakes, but today is also the worst time for them in the future. With each model release, they will only get better. Just like self-driving cars, five years ago, they might not have been safe for humans, but today it’s 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 use.

5. 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 lies now?

Haseeb: I think the most important aspect is Privacy. AI itself is actually very dangerous to privacy because every time you give data to AI, it usually means giving data to a third party. Psychologically, we feel that our conversations with AI are very private, like an extension of our brain, but from a technical and legal perspective, it doesn’t necessarily belong to your own private space.

Haseeb: AI will also make large-scale surveillance much easier. Data that was previously scattered in different places can now be instantly integrated, analyzed, and even de-anonymized by models. Therefore, as AI enhances the capabilities of states and large institutions, there needs to be another set of technologies to enhance individual capabilities. This is the significance of privacy-preserving cryptography, Zcash, Monero, and private AI infrastructure.

Haseeb: Venice is a great example. You can certainly use traditional Login and Credit Card, but if you really want strong privacy, you can use Crypto Wallet, make Crypto Payments, and combine it with Encrypted Inference. If you use a Credit Card, the platform will definitely know who you are. But if you use Crypto, along with true End-to-end Encryption, the platform might not even know your name, location, or what Model you are using.

Haseeb: So now I will say something very different from what I said at the beginning of the year: Crypto is not for your AI Agent, but for you. Your Agent is just there to use it on your behalf.

6. The First to Be Eliminated by Agentic Finance is the "Lazy Tax" of Banks and Brokers

Mr. Z: You’ve been discussing Agentic Bank Run recently. If everyone’s Agents start managing money automatically, will it make the financial system more fragile?

Haseeb: What happens in the first phase might not be a Bank Run at all, but many Business Models built on human laziness will be destroyed. Suppose you have a lot of Cash in a JPMorgan Checking Account with a very low interest rate, while Money Market could give you much higher Yield. Why doesn’t everyone move their money to the highest yielding place every day and then move it back when they need to pay? The answer is simple: because people are lazy.

Haseeb: Agents won’t be lazy. They can check your funds 24/7, moving Idle Cash to the optimal place automatically. Cheap Deposits are a core part of many large banks' business models because consumers are too lazy to move their money, allowing banks to lend at very cheap rates and earn a Spread. Many Brokers are the same; a lot of Schwab and Fidelity's economics come from customers keeping Cash in Sweep Accounts.

Haseeb: Agents will arbitrage this "human laziness tax". What is most likely to happen is not that all banks will suddenly be drained, but that banks will be forced to raise Deposit Rates to retain funds. The result is that Funding will no longer be so Cheap, compressing the Margin and ROE of banks and Brokers.

7. Amazon Has Already Started Blocking Agents: Agent-first and Local Agents Will Win

Mr. Z: But these Incumbents won’t just sit back and let Agents take their profits, right?

Haseeb: Of course not. The second path is to directly Block Agents. This is already starting to happen. If a platform's profits largely come from Advertising, Sponsored Placement, or User Friction, a true Agent representing user interests is a threat. Because Agents don’t care about what is a Sponsored Result; they will only look for the best results for users.

Haseeb: Banks will have the same Incentive in the future. They won’t say, "I don’t want your Agent to move Deposits to higher Yield places." They will say, "It’s not safe," "There’s Compliance Risk," or "AI might Hallucinate." But the underlying conflict of interest is essentially the same.

Haseeb: So in the long term, I am particularly optimistic about two types of things. The first is Agent-first Companies that are willing to embrace Agentic Commerce from day one. The second is Local Agents. If the Agent you are using essentially belongs to Meta or another third-party platform, they can Block that company; but if AI truly runs on your own computer, operating through a Browser like a human, then it will be very hard to Block you.

Haseeb: Ultimately, the biggest transfer of power is from "the platform deciding what you can do" to "your Agent truly acting in your interests."

8. 2026 Prediction Review: BTC 150k is Difficult, but Fintech Chains Are Indeed Overvalued

Mr. Z: At the end of last year, you made a series of 2026 Predictions. Looking back now, which judgments were the most accurate, and which need correction?

Haseeb: Overall, I think it’s not bad, but of course, there are mistakes. I predicted BTC would reach $150,000 by the end of the year. Looking at the current position, it would need to nearly double to reach $150,000 by year-end, which is very difficult. However, I think returning to above $100,000 is still possible.

Haseeb: Regarding Fintech Chains, I think the direction is basically correct. Last year, the market believed that Tempo, various Fintech Chains, and Robinhood Chain would absorb a lot of Stablecoins and RWA through Distribution, even challenging Ethereum and Solana. But so far, the scale has not appeared as everyone imagined.

Haseeb: Robinhood is still early, so I won’t draw direct conclusions. But the real significant increase in RWA is still mainly happening on Ethereum and Solana. Distribution is important, but Distribution itself does not automatically create a Blockchain Ecosystem.

9. Robinhood is Not the Next Solana: What’s Really Lacking is RWA and Developers

Mr. Z: I want to challenge this point. Robinhood Chain has recently introduced Agent Finance, Programmable Hooks, Stock Token Liquidity, Perps, and some new Market Structures. Do you still think it is far from Solana?

Haseeb: I think Robinhood Chain is still TBD (To Be Determined). There is indeed a lot of Meme Activity now, but Robinhood was not meant to be a Meme Chain; it is really about RWA. And currently, the real RWA Activity on-chain is still very small, accounting for a limited proportion of the entire RWA market.

Haseeb: Another issue is Developers. I currently do not see many truly independent developers continuously building new Applications and Ideas on Robinhood. Right now, it’s more about Meme Farming, Meme Cycle, and some Market Structure Experiments. Of course, Perps and DeFi Market Structures are worth paying attention to, but in absolute scale comparison, Robinhood Chain is still much smaller than Hyperliquid, Arbitrum, BSC, Tron, Base, and Solana.

Haseeb: I’m not saying it won’t succeed. I just feel that the market previously directly inferred that "Robinhood has Distribution" means "Robinhood Chain will swallow everything," and this has not been proven today. What is truly worth tracking is whether it can show 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 Only Be 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 as far from what I originally imagined. In today’s 24-hour Snapshot, Hyperliquid Reported Volume is about $4 billion, while Variational (@variational_io) and Lighter (@Lighter_xyz) each have about $1.5 billion. Of course, it’s not yet the 40 / 30 / 20 I originally predicted, but the structure of a few large Venues is starting to appear.

Haseeb: We are investors in Variational and Lighter and hold some HYPE, so I need to disclose this. But I still believe that as the Liquidity Flywheel of these Venues matures, the market structure will continue to Consolidate.

Haseeb: Another thing I am quite sure about is Equity Perps. At the beginning of the year, this market was basically close to 0, but it has become significant now. I think one important growth curve for future DeFi Derivatives is to bring stocks and more Traditional Assets into the same Perpetual Infrastructure.

11. HIP-3 is Not as Permissionless as Imagined: TradeXYZ (@tradexyz) is More Like a Deep JV

Mr. Z: I personally really like HIP-3. Jeff essentially tells everyone that you can come to Hyperliquid to open your own Casino and set your own rules as long as you pay Rent. Why are you a bit Skeptical about this Model?

Haseeb: Because if the HIP-3 Model were inherently so effective, we should have already seen many successful cases before TradeXYZ appeared. But that’s not the case. TradeXYZ is indeed a huge success, but it is not a random team that walked in through the Front Door and automatically succeeded because of HIP-3.

Haseeb: It has a deep history with the Hyperliquid Ecosystem, including previous collaborations in areas like Spot Infrastructure. So this feels more like a deep partnership, even a quasi-JV, rather than proving that a completely neutral permissionless marketplace can grow great companies on its own.

Haseeb: If you really want to build a very large market, you will encounter a very practical problem: why should I build my entire company on someone else's land, paying a lot of rent in the long term, while accepting the platform risk that the other party can change the rules in the future? If you are the next Polymarket, you would likely want to own your own infrastructure.

Haseeb: Hyperliquid can provide you with good rails, but it won't automatically do market development for you. Building an entirely new category from 0 to 1 still requires founders, a complete team, capital, and distribution. Protocol primitives can lower execution difficulty, but they cannot replace company building.

12. Hyperliquid is Not Eating CEX: It is Actually Creating a New Market

Mr. Z: The most common saying in the market is that Hyperliquid is stealing market share from Binance, OKX, Bybit, and Bitget. You think this statement is not entirely correct?

Haseeb: If you really analyze the addresses on Hyperliquid and compare them with CEX usage, you will find that the overlap between users on both sides is much lower than people think.

Haseeb: Three to four years ago, there were still many non-KYC CEXs globally, and the user base of those exchanges might be closer to today's Hyperliquid. But now all major CEXs are highly regulated and basically require KYC. A group of users, due to region, KYC, regulation, or other reasons, originally did not have good access to CEXs. For them, Hyperliquid is not "moving from Binance" but finally having a place to trade.

Haseeb: If Hyperliquid were really eroding CEX, you would see Hyperliquid volume rising while CEX volume declines. But what you see now is growth on both sides. So I think it is closer to market expansion rather than market share transfer.

13. HYPE's Next Growth Curve: Pre-IPO Has PMF, but Crypto Core Remains the Foundation

Mr. Z: The traditional market will eventually get closer to 24/7 trading. When "around-the-clock trading" is no longer a moat, where will Hyperliquid's next growth curve be? Pre-IPO, private markets, HIP-3?

Haseeb: Pre-IPO equity clearly has demand and product-market fit. TradeXYZ still has a lot of room for further expansion. But if you ask me what Hyperliquid's next biggest product curve is, I don't have a very certain answer right now.

Haseeb: However, I think sometimes people are too obsessed with "how Hyperliquid can step out of crypto," and instead overlook that Hypercore's own crypto business is very strong. I am relatively optimistic about the crypto market in the next 12 months, so if crypto continues to recover, I think HYPE itself, as well as exchanges built around crypto trading, will still have a good environment.

Haseeb: The ideal outcome is not for Hyperliquid to leave crypto, but for the crypto core to continue compounding, while equity, pre-IPO, and other financial assets layer by layer expand the TAM.

14. Dragonfly is No Longer Just a "Bridge Between East and West": The Next Stage Looks at the Complete Global Crypto

Mr. Z: Dragonfly's initial important positioning was to connect China's first generation of internet capital with global crypto, led by Bo Feng (@FengBo22588938). Today, the Sino-US environment is completely different; should Dragonfly still be understood as an East-West bridge?

Haseeb: When I first entered crypto, China was relatively constructive towards the entire industry. In 2018, there was certainly an ICO crackdown, but crypto did not completely hide away. By 2020, when the market entered a rapid growth phase, there was also a sense that crypto would ultimately become a globally accepted technology.

Haseeb: After 2021, when China banned crypto and mining, many founders who originally built in China moved to Singapore, Hong Kong, and later to Dubai and the US. So today, I wouldn't say that Dragonfly's work is just about connecting China and the US. Crypto itself is global.

Haseeb: The Chinese diaspora is still a very important part of this industry; they have just scattered across different countries. Our LPs include Asian capital and US institutions. I often say that if you really want to do well in crypto investing, you must see the whole elephant, not just touch it blindly.

Haseeb: This is also why we have layouts in both Singapore and New York. We want to see founders who are genuinely creating value from different parts of the world, rather than deciding whether to invest based on where the company is registered. I have long believed that China will eventually have a moment of reawakening, realizing that crypto has grown too large to be completely ignored. It is too important and resembles a foundational technology of the 21st century.

15. Crypto VC Will "Disappear": By 2030, Only Those Who Truly Invest Will Remain

Mr. Z: You previously mentioned that as crypto matures, "crypto VCs" may disappear in the future. By 2030, what kind of funds will still qualify to exist?

Haseeb: I am not saying that all crypto investment firms will disappear, but we will no longer call them crypto funds. Crypto has already entered the water supply. Stablecoins will become the fintech backend, neobanks will integrate crypto, and consumer products will have built-in wallets; 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, this classification made sense. But later, all companies became internet companies, and there was no longer such a thing as internet investing, just investing. I believe crypto will eventually reach the same state.

Haseeb: This transition will be very brutal for those 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 call yourself a crypto VC. It will become increasingly difficult in the future.

Haseeb: The real durable edge will become: do you really understand this world? Do you have real relationships? Can you help founders solve problems? Besides "do you understand crypto," what else can you bring? If you have nothing, you will not survive this transition. If you do, I actually think there will be great opportunities in the future.

Mr. Z: So now that you are in Asia, who do you most want to meet: researchers, investors, or others?

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: To conclude this episode, I think we can summarize it into four judgments. First, AI agents have not made crypto the default payment rail as the market imagined at the beginning of the year; cards have quickly captured a large share of consumer payments. Second, the real opportunities for AI and crypto may be even greater; agents will rewrite crypto UX, and privacy will become an important infrastructure for protecting personal autonomy in the AI era. Third, Hyperliquid is currently more about creating a new market rather than simply stealing volume from CEXs, and HIP-3 itself will not automatically create a great company. Fourth, crypto is transitioning from an independent industry to the underlying infrastructure of the entire financial and tech world, and even the classification of "crypto VC" may lose its meaning in the end. Thank you, Haseeb, for joining us again at 168X!

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.