Dragonfly Managing Partner Haseeb Qureshi is raising concerns about the rapid timelines projected for agentic payments in the cryptocurrency sector. Dragonfly is a prominent global investment firm focused on decentralized technologies and digital assets, with Haseeb Qureshi recognized for his in-depth analyses and active engagement in crypto industry debates. His recent comments reflect a more cautious stance compared to the widespread optimism among crypto leaders and payments specialists regarding the near-term prospects of machine-led or agentic transactions.
Historic Parallels And Present-day Challenges
Qureshi draws parallels between the current stage of agentic payments and the early period of innovations like the computer mouse, noting that significant technological advancements can remain niche for decades before achieving widespread use. Despite the enthusiasm around projects such as OpenClaw, he points out substantial technical limitations in current agentic payment tools.
He describes OpenClaw as unreliable and complex, highlighting issues with bugs and inadequate financial decision-making. The crux of these difficulties, he explains, is that core models struggle with tasks that fall outside their original training data, resulting in unpredictable or flawed outcomes when handling real money.
Training Data And Next-generation Models
Haseeb Qureshi emphasizes that a key factor holding back progress is the lack of reinforcement learning based on real-world agent interaction data. While ample data now exists across leading AI research institutions, none have yet applied this specific training approach to live agentic payment systems such as OpenClaw.
He anticipates that integrating this data could significantly boost agent performance and reliability. Qureshi forecasts the arrival of next-generation models in the coming months, which may transition agentic payments beyond the experimental or “tinkerer” phase and encourage adoption by early technology users. However, he remains convinced that mainstream usage is still several years away.
To support his view on the technology’s maturity, Qureshi references current usage statistics. The x402 agent, for example, processes about $1 million in daily transaction volume, and the machine payments protocol (MPP) remains even lower. These figures, he argues, reinforce the experimental status of agentic payment methods at this stage.
Agentic payments will be a huge trend. But it’s not here yet. I had the most bearish take in this article, and I want to explain why, because crypto has a bad habit of getting high on its own supply and giving unrealistic timelines about technology.
Long-term Outlook And Cautions For The Crypto Sector
Qureshi remains optimistic about the eventual impact of agentic economies, repeating his expectation that agents will transform global money movement. According to his perspective, innovation in this field will likely follow classic adoption patterns, moving from initial enthusiasts to a broader early majority and later, mass market participants.
He cites investor Chris Dixon’s view that technology adopted by enthusiasts during late hours eventually goes mainstream within a decade. For Qureshi, agentic payments fit this pattern, and he maintains that the core technological direction is sound.
Despite his belief in the field’s potential, he cautions the crypto industry to maintain realistic expectations. Qureshi argues that patience is essential and that overhyping rapid deployment timelines could undermine the sector’s credibility as agentic payment applications develop.
