The debate over the readiness of autonomous AI agents to manage cryptocurrency capital has intensified, with prominent industry figures expressing starkly different views on timing and risk tolerance. At NEARCON 2026, a sharp disagreement emerged between Dragonfly Managing Partner Haseeb Qureshi and Kraken co-CEO Arjun Sethi.
The core of the debate centered not on if AI agents will eventually manage capital—both agree they will—but on how soon they can be trusted with real money. Qureshi expressed significant skepticism, arguing that current technology is not sufficiently reliable for meaningful economic activity. "Something that works with money 90% of the time is unusable for actual economic activity," Qureshi stated, adding that even 95% reliability is insufficient. He characterized the current state as being in the "nothing phase" of a progression toward widespread adoption, cautioning against extrapolating from viral social media demos.
"You want to be very cautious of trying to ingest your worldview of technology by reading Twitter hype people and watching Twitter demos," Qureshi warned. He bluntly stated that for major consumer platforms, deploying such systems prematurely is not viable: "You cannot do that sh**."
In contrast, Arjun Sethi of Kraken presented an optimistic, exponential view of innovation. "We think we know what's going to happen. The speed and the level of innovation... is exponential," Sethi argued. He revealed that Kraken is already developing agent-like capabilities for customers, suggesting deployment is "weeks and months away — not years away." Sethi believes defensive security capabilities will scale alongside the growing attack surface.
The divergence was crystallized in a rapid-fire question session. When asked what percentage of his own portfolio an AI could manage better today, Qureshi cautiously answered "Five percent." Sethi's response was definitive: "One hundred." Pressed on whether he would entrust all his crypto to an autonomous agent within a year, Sethi confirmed, "Everything. In the next six to twelve months."
This debate reflects a broader industry split on whether autonomous finance is a near-term inevitability or remains a frontier experiment. Parallel commentary highlights the urgent need for infrastructure to support this shift. An opinion piece argues that as AI agents become economic actors—with Gartner projecting over 40% of enterprise workflows involving them in 2026—the lack of standardized on-chain identity poses a systemic risk.
The article emphasizes that traditional API keys and credentials are inadequate for autonomous decision-makers. It proposes blockchain-based "Know Your Agent" (KYA) frameworks as a crucial trust layer to verify identity, mandate, and accountability. The piece warns that without such verifiable, programmable identity infrastructure, trust in AI-driven finance will fracture, potentially leading to regulatory backlash that could slow the adoption of both tokenization and autonomous finance. The convergence of AI and tokenization is creating a market structure where machine-driven actors may soon outnumber human traders in certain domains, making this infrastructure debate critical for the industry's future.
