Dragonfly Executive Haseeb Qureshi: AI Agents Don't Need Cryptocurrency, What They Need is Us Humans

Coin Circle (120btc.CoM): Haseeb Qureshi, managing partner of the well-known crypto venture capital firm Dragonfly, publicly corrected his previous industry assessments. At the TOKEN2049 Singapore summit held on October 7, Beijing time, he delivered a speech titled "Agents and Crypto: What's Real and What's Cope," clearly denying the early assumption that "AI agents will fully adopt cryptocurrency." On October 19, Beijing time, he reposted the official speech video on social media platform X and reiterated, "Agent programs do not need cryptocurrency; what truly needs it is us humans."

Earlier this year, Binance founder Changpeng Zhao and Coinbase CEO Brian Armstrong both predicted that agent payments were about to explode, and the trading volume of the x402 protocol led by Coinbase once soared, with the market generally anticipating that AI would replace humans as the core audience for digital assets. However, real data has poured cold water on this narrative. Qureshi bluntly stated that up to 90% of the payment flow on the x402 protocol was purely for the purpose of gaming incentives. Currently, AI agents aimed at the public still prefer traditional credit cards for actual settlements. Financial institutions, including Rain, Visa, Ramp, and Stripe, have launched sub-cards specifically for agent use, equipped with precise limits and consumption scenario restrictions. Qureshi revealed that he had configured both a cryptocurrency wallet and a traditional credit card for his AI agent, but the vast majority of purchases were completed via card swipes, primarily due to the current mainstream commercial environment not widely accepting cryptocurrency settlements. Although cryptocurrency assets still have advantages in specific scenarios like cross-border payments, traditional financial channels are actually well-prepared to accommodate AI payments.

AI Agents Will Evolve into Intelligent Risk Control Intermediaries

After stripping away the halo of payment media, Qureshi believes that the core reshaping of the crypto ecosystem by AI agents lies in "simplifying complexity." For a long time, the underlying interaction logic of cryptocurrencies has been extremely obscure, with ordinary users frequently facing wallet management, high network fees, cross-chain friction, and security traps such as phishing attacks or honeypot contracts. On-chain data estimates that about 20% of Bitcoin in the entire network is likely to be permanently lost due to poor private key management.

In the traditional financial system, commercial banks act as an intelligent risk control layer between users and funds, capable of timely intercepting abnormal instructions such as "remitting to high-risk areas illegally." In contrast, the crypto world lacks such fault tolerance mechanisms. Qureshi pointed out that when AI agents fully take over the front-end on-chain interactions for users, they will perfectly fill this risk control gap. These intelligent programs can automatically verify the status of target addresses before executing transfers, keenly capture anomalies when encountering Domain Name System (DNS) hijacking, and pre-check whether smart contracts are under hacker attack before authorizing calls. Due to the tireless and diligent nature of algorithmic programs, they will become the most solid firewall for users navigating the on-chain ecosystem.

Minimalist Privacy Tools Against Centralized Surveillance

The latter half of the speech focused on data privacy. Qureshi proposed that absolute privacy protection does not exist in reality; it is essentially just a "rate limit," relying on extremely high time and physical costs to infringe upon privacy. He cited U.S. Supreme Court Justice Samuel Alito's viewpoint, stating that in the pre-computer era, the greatest umbrella of privacy protection was not legal texts but the real-world resistance of physical tracking. However, the assertion by Anthropic CEO Dario Amodei indicates that powerful AI computing power can automate the assembly of a complete personal life trajectory at a very low cost.

Taking the controversial U.S. security surveillance company Flock as an example, its centralized database can retrieve the underlying footage of about 80,000 cameras with a single query. This scale of monitoring capability also threatens the life and property safety of digital asset holders. In 2025 and 2026, cases of real physical attacks against crypto investors have reached new highs. Criminals only need to feed leaked fragmented data to large language models to quickly pinpoint target addresses and fund amounts.

Qureshi concluded that if AI is a technology that reinforces centralized surveillance, then cryptocurrency assets are tools that return data control to individuals. In the future, AI agents will lower the threshold for privacy protection to the extreme—"privacy is merely a prompt." Users can simply issue commands to require agents to erase public network footprints, simulate hacker offense and defense tests, or request to bypass transparent ledgers by directly using privacy networks like Zcash during transfers. As part of Dragonfly's investment landscape, the privacy-focused AI platform Venice allows users to log in without a password using only their Ethereum address and settle using stablecoins, achieving absolute identity concealment.

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.