Dragonfly Partner Haseeb: Cryptocurrency is not designed for humans; it is built for AI intelligence.

Dragonfly Partner Haseeb Qureshi made a bold statement on the Bankless podcast: Cryptocurrency was never designed for humans—the ultimate users are AI intelligences. This article is based on an interview from Bankless, edited and summarized by Dongqu.
(Background: Dragonfly Managing Partner Haseeb Qureshi says “AI is not inherently radical”! Despite the bear market, they raised $650 million)
(Additional context: Don’t blindly follow OpenClaw. The AI lobster is strong, but not necessarily suitable for you)

Table of Contents

Toggle

• Crypto’s “friendly fire” everywhere, but AI won’t get hit
• The promise of smart contracts, destined to be fulfilled by AI
• The future of finance: just speak, and AI will handle it
• Dual worlds: safety camp vs. frontier camp
• Dragonfly’s $650 million new fund, heavily focused on AI × Crypto

Have you ever wondered why, after ten years in crypto, you still sweat when signing large transactions? Haseeb Qureshi believes it’s not your fault—it’s because you’re not the “target user” of this technology. In the latest Bankless podcast, he presented a provocative idea: blockchain might have been built from the start as infrastructure for AI intelligences.

Crypto’s “friendly fire” everywhere, but AI won’t get hit

Haseeb bluntly states that the crypto world is full of “anti-human” designs: address poisoning attacks, expired authorizations, phishing URLs, blind signing—these are nightmare-level traps for humans, but for AI intelligences, they’re just code to parse line by line.

“I never worry about a wire transfer accidentally sending money to North Korea,” Haseeb says, “but every time I sign a crypto transaction, I have that thought.”

He points out that the crypto community has long shifted responsibility onto users—“you should be more careful,” “develop better habits”—but if these issues persist after ten years, maybe it’s not the users who are at fault, but that we’ve chosen the wrong users.

The promise of smart contracts, destined to be fulfilled by AI

Haseeb recalls his early ideals: “Smart contracts should replace legal contracts, no need for lawyers, just sign agreements with code.” But the reality is, even Dragonfly, as a top crypto VC, still requires legal contracts before buying tokens—smart contracts are just “additional insurance.”

Where’s the problem? Legal contracts are full of randomness—uncertain jurisdictions, randomly selected judges and juries, clauses that might be invalid. But human intuition still sees legal contracts as “more predictable” than smart contracts.

“This is because of our limited rationality,” Haseeb explains,

Humans are far less capable of processing code than AI. For AI, smart contracts are just step-by-step analyzable EVM bytecode, with 100% certainty. The promises crypto initially made—better enforcement, better property rights—can only truly be realized by AI intelligences.

The future of finance: just speak, and AI will handle it

Haseeb envisions a near future: you no longer need to open DeFi protocol websites, compare TVL, or research contract security. You just tell AI, “I think interest rates will rise, help me reduce risk,” and the AI will automatically filter protocols and execute the best strategies.

“In this world, you won’t click on protocol logos, won’t look at their marketing, and might not even specify which protocol to enter,” he says. This means many current business models—brand recognition, network effects—will be replaced by AI’s rational filtering.

The ultimate beneficiaries will be consumers, as efficiency is captured by users.

Dual worlds: safety camp vs. frontier camp

However, Haseeb admits that full AI takeover of crypto operations won’t happen overnight. He predicts a “dual-track” future:

Safety track—big players like OpenAI, Google will take a conservative approach, requiring human approval for each transaction. Legal liability and brand risk prevent them from fully releasing control. “If Claude messes up a trade and causes users to lose money, Anthropic can’t afford that headline,” Haseeb says.

Frontier track—open-source AI projects like OpenClaw, unbound by legal liabilities, with risks borne by users. They prefer command-line interfaces, raw data, direct private key operations—skipping all human-designed UI. Haseeb compares this to early automobiles:

Ten years from now, we’ll look back in horror, wondering how we let humans manually sign transactions blindly, check addresses visually, and rely on Twitter to verify protocols weren’t hacked.

Dragonfly’s $650 million new fund, heavily focused on AI × Crypto

Regarding Dragonfly’s latest $650 million fund, Haseeb states the team is “focusing heavily” on the intersection of AI and crypto, but he also remains pragmatic: “AI intelligences are general intelligences. They use the same tools we do, so there might not be many investments specifically targeting AI. But if you believe in the AI agency theory, just like believing in China’s crypto ban being lifted—everything will go up, and overall demand growth is bullish for crypto.”