At Consensus Hong Kong this week, top crypto investors weren’t packing up — they were recalibrating. After a prolonged market downturn that cooled speculative excess, venture capitalists described the current moment not as retreat but as selective redeployment: double down on proven winners and pick a few high-risk, potentially transformative plays. Dragonfly managing partner Hasseeb Qureshi captured that stance as a “barbell” approach. On one end sit durable, revenue-generating verticals — stablecoins, payments and tokenization — where product-market fit is clear and scaling makes sense. “There’s stuff that’s working, and it’s just like, scale it up, go even bigger,” Qureshi said. On the other end are far risker, next-gen bets, chief among them crypto’s convergence with artificial intelligence. Qureshi said he’s exploring AI agents that can transact onchain, even while conceding the real-world fragility of those early systems (“if you give an AI agent some crypto, it’s probably going to lose it within a couple days”). The opportunity is substantial, but so are attack vectors and design flaws — meaning careful vetting and technical depth are prerequisites. That caution is tempered by hard-learned lessons. Qureshi admitted he initially wrote off NFTs as “definitely a bubble,” only to pivot and back infrastructure plays such as Blur months later. Dragonfly also famously missed an early chance at Polymarket: Qureshi said the fund issued the first term sheet to founder Shayne Coplan but passed when a rival offered a higher valuation. “Generational miss,” he called it — though Dragonfly later joined a 2024 round ahead of the U.S. election and is now a major shareholder. The anecdote underlines a recurring theme at the conference: deep thematic conviction can take years to pay off. Mo Shaikh of Maximum Frequency Ventures echoed that long-horizon view. His most successful thesis wasn’t a short trade, he said, but a 15-year bet that blockchain could fundamentally re-architect financial risk systems. “Have a 15-year timeline,” Shaikh urged, warning founders and investors against 18-month cycle thinking. Data from Pantera Capital’s Paul Veradittakit supports the narrative that capital has concentrated into fewer, higher-quality opportunities. Pantera’s managing partner said crypto VC capital rose 14% year over year even as deal count fell 42% — a “flight to quality” as investors back accomplished teams and tangible use cases rather than hype. Veradittakit also noted the institutionalization of crypto fundraising: what began as $25 million early funds dominated by family offices has grown into a roughly $6 billion platform, with institutions increasingly driving the next leg of capital formation. His blunt advice to founders in a softer market: focus on product-market fit. “If there is a token, it’ll naturally come,” he said. The message from Consensus Hong Kong was straightforward: in a downshifted cycle, scale what demonstrably works, be surgical with speculative bets (especially in nascent AI-on-chain use cases), and don’t mistake narrative momentum for fundamentals. For many VCs, success now means patience and a willingness to play a 15-year game. Read more AI-generated news on: undefined/news