Dragonfly managing partner Haseeb Qureshi used his TOKEN2049 appearance in Singapore on Oct. 7 to revise a thesis he had previously pushed on AI agents and crypto. In a talk titled Agents + Crypto: What's Real & What's Cope, Qureshi said his earlier call was wrong: agents do not necessarily need crypto. Users do.
The speech, published in full after being organized by ChainCatcher, opened with a direct admission. Qureshi said he had argued for some time that agents would use crypto and that traditional financial rails would not be ready before agents arrived at scale. That is not what happened.
Cards, not crypto, are handling most consumer agent payments
Qureshi revisited the narrative that circulated earlier this year, when many in crypto argued that agentic payments were about to take off and that AI, not ordinary people, would become crypto’s real end user. He said that story did not play out.
As one example, he pointed to x402, Coinbase’s agent payment protocol. According to Qureshi, agents have not adopted it at scale, and 90% of payment volume on x402 has come from farming activity.
At the same time, he said agents are already here. Consumer agent products are growing quickly, and OpenAI is also building in the area. The key question, in his view, is what payment rails those products are actually using right now.
His answer was simple: cards, especially credit cards.
Qureshi said companies including Rain, Visa, Ramp and Stripe have already launched agentic card products. These products create sub-cards under a user’s main card, with spending limits and merchant or use-case restrictions. He argued that these firms have recognized agentic commerce as a major opportunity and have already prepared for it.
He also used his own setup as an example. He said he had given an agent both a crypto wallet and a card. In practice, most of the things the agent bought for him were paid for with the card.
That does not mean cards solve every payment problem. Qureshi noted that cards are nearly everywhere in developed markets such as the United States, but not in every part of the world, while agents will be widespread. Even so, he said the early-year claim that agents would use crypto because traditional rails would not be ready has turned out to be wrong. The rails are already in place.
The overlap between agents and crypto is narrower than the market story suggested
Qureshi described part of the earlier thesis as cope. Outside standard use cases such as international payments and large transfers, he said, the overlap between agents and crypto is not as broad as many had claimed at the start of the year. In his framing, the things agents are likely to do with crypto look a lot like the things people already do with crypto today.
Still, he did not argue that agents are irrelevant to the sector. His point was that the real impact lies elsewhere.
Agents make hard things easy, and governments are feeling it
For Qureshi, the most important property of agents is that they make difficult tasks easy. He first applied that idea to public services.
Governments around the world, he said, are already feeling pressure because many public systems rely on bureaucratic friction. Filing court claims, submitting complaints to the U.S. Consumer Financial Protection Bureau, applying for benefits, or lodging housing complaints all involve layers of process that ordinary people often struggle to get through.
That friction is now breaking down. Qureshi said application volumes are surging across many such services because AI agents can faithfully act on behalf of their owners and consume public-service capacity. In his telling, governments are being flooded by AI agents.
The deeper issue, he said, is that these systems were never designed for a world in which everyone could easily file every complaint or claim. They functioned as a hidden rationing mechanism, limiting access through hassle and complexity. In an agent-driven environment, that mechanism stops working.
Why he thinks agents could make crypto more reliable
Qureshi then turned to crypto itself. He argued that one of crypto’s biggest barriers has always been difficulty. Many users do not know how to operate wallets, manage gas fees, or use bridges. They are not only intimidated; they also make costly mistakes. He cited hacks, phishing and honeypots, and said on-chain records show that a meaningful share of assets has been permanently lost through errors of that kind.
That is where he sees one of the biggest changes agents could bring to crypto: making it more reliable by protecting users from their own mistakes.
He compared the situation with banking. Even after more than a decade in crypto, he said, sending a large transfer from a wallet can still feel risky. Bank transfers do not create the same anxiety because a bank sits between the user and the money as an intelligent layer that can block obvious errors.
Qureshi argued that once AI becomes the API through which people access crypto, a similar protective layer can emerge on-chain. In that setup, agents could screen for phishing, stop users from landing on fake interfaces after DNS poisoning, check contracts before interaction, and verify destination addresses before funds are sent. His claim was straightforward: agents do not get lazy, skip steps or get tired, so they do not make the same mistakes people do.
On that basis, he said the way people use crypto five years from now will not look like the way they use it today, and the trust relationship between users and crypto will be reshaped.
Privacy, in his view, has always depended on friction
The second half of the talk focused on privacy. Qureshi’s argument was blunt: privacy has never really functioned as a right. It has functioned more like a rationing system sustained by real-world friction.
He cited a U.S. Supreme Court justice in a GPS tracking case, saying that in the pre-computer era the strongest protection for privacy was not constitutional or statutory but practical. Privacy existed in part because police did not have the time or resources to follow every suspect around the clock.
Technology, he said, has changed that balance. He also referenced Anthropic founder Dario Amodei and his clash with the U.S. Department of Defense, saying the same point had surfaced there: powerful AI makes it possible to automatically assemble a detailed picture of anyone’s life at scale. Personal data is already scattered across leaks, he said. What used to stop others from reconstructing a full profile was cost.
From Flock cameras to public crypto holdings, scale changes the equation
Qureshi used the controversy around U.S. camera company Flock as an example. Cameras on street corners are not new and are not illegal, he said. The difference is scale. When every corner has a camera and AI can search all footage instantly, quantity becomes a qualitative shift. Government can suddenly know where everyone is, all the time.
He said crypto is seeing a parallel development. Physical attacks against crypto holders rose sharply in 2025 and 2026, with reported incidents reaching an all-time high. For people whose holdings are publicly visible, the old comfort was that no one would spend the effort to target a small individual. Qureshi argued that this assumption is weakening because attackers no longer need the same level of manual research.
AI strengthens the state; crypto can strengthen the individual
He framed this as a panopticon problem and said it is one of the main reasons crypto matters. In his view, AI and crypto push in opposite directions on this axis. AI strengthens the state by removing old constraints: slowness, limited analytical capacity and poor data processing. Crypto pushes power back toward the individual.
In a world where governments become stronger and more capable of seeing everything, he said, crypto is one of the few available hedges for individuals.
He also acknowledged a longstanding problem for privacy technology: people say they care about privacy, but they routinely hand over chats and data to Meta, ChatGPT and Google. Privacy wins only when it becomes easy, he argued, not simply because it is important.
To make that point, he pointed to several historical examples. Encrypted web traffic became the default only after Let’s Encrypt made HTTPS certificates close to free. Before that, SSL was slow and expensive and was often limited to checkout pages. Messaging followed a similar path: only when Signal became usable and WhatsApp rolled out end-to-end encryption broadly did encrypted communication reach billions of users. He said Zcash is now showing a similar pattern, with better user experience directly explaining growth in shielded address usage.
His bet: AI agents could make privacy tools usable at near-zero cost
Qureshi’s core prediction is that AI will drive the same kind of shift for privacy. A sufficiently capable personal agent, he said, could act as a privacy defender. A user might simply say, “Make me invisible,” and the agent could remove public traces, improve operational security, scrub information from the dark web, and model how an attacker might identify or target that person.
In that world, privacy protocols such as Zcash may remain difficult under current interfaces, but the difficulty would matter less if agents become the user’s interface to crypto. The agent could decide when not to use Bitcoin, when to use Zcash instead, and when plaintext should be replaced with encrypted text.
The final piece is private AI
Qureshi said the last piece of the picture is private AI. He cited Venice, a privacy-focused AI platform in Dragonfly’s portfolio, as an example. He described it as a private version of ChatGPT that runs on crypto rails. Users, he said, can even use it anonymously by registering with an Ethereum address and paying in stablecoins.
He argued that this will be one of the most important civil-liberties battles of the next decade. The questions, as he framed them, are how individuals protect themselves against expanding state power and how they retain control over AI. AI, he said, will be the most politicized technology of the next 10 years.
His closing line: users need crypto more than agents do
Qureshi ended where he began. The story that agents will use crypto for everything they do does not hold up, he said. For most consumers, agents will use many of the same tools people already use. But his final conclusion was clear: the real need for crypto in the agent era belongs less to the agents themselves than to the people behind them.