The mood among top venture capitalists at Consensus Hong Kong was not retreat, but recalibration, as the crypto market experienced a prolonged downturn.

Hasseeb Qureshi, managing partner at Dragonfly, described today’s venture market as a “barbell:” On one side, proven verticals compounding at scale; on the other, a narrow set of high-risk, next-generation bets.

“There’s stuff that’s working, and it’s just like, scale it up, go even bigger,” Qureshi said, pointing to “stablecoins, payments and tokenization in particular.” In a market that’s cooled from speculative excess, these are the sectors still demonstrating product-market fit and revenue.

On the other side is crypto’s intersection with artificial intelligence (AI). Qureshi said he is spending time on AI agents capable of transacting onchain, even though if “you give an AI agent some crypto, it’s probably going to lose it within a couple days.” The opportunity is real, but so are the attack vectors and design flaws.

The cautious tone reflects lessons learned. Qureshi said he initially dismissed non-fungible tokens (NFTs) as “definitely a bubble,” only to reverse course months later and back infrastructure plays like Blur. That experience, he said, was a reminder to balance conviction with adaptability in fast-moving cycles.

Dragonfly also famously missed an early opportunity in prediction market Polymarket.

“We were actually his first term sheet,” Qureshi said of founder Shayne Coplan, but passed when a rival fund offered a higher valuation. “Generational miss,” he called it, although Dragonfly later joined a 2024 round before the U.S. election and is now a major shareholder. The takeaway: Thematic conviction, in this case around prediction markets, can take years to pay off.

Maximum Frequency Ventures’ Mo Shaikh argued that venture success in crypto still hinges on long time horizons. His best thesis, he said, wasn’t a trade but a 15-year bet that blockchain could re-architect financial risk systems.

“Have a 15-year timeline,” he advised, urging founders and investors to resist 18-month cycle thinking.

If the venture environment feels tighter, Pantera Capital’s data supports it. Managing partner Paul Veradittakit said crypto VC capital rose 14% year over year, even as deal count fell 42%, evidence, he said, of a “flight to quality.” Investors are concentrating into “accomplished entrepreneurs” and “tangible use cases.”

After more than a decade fundraising in crypto — from $25 million early funds dominated by family offices to today’s $6 billion platform — Veradittakit sees institutions increasingly driving the next leg. But his advice to founders in a softer market was blunt. “Focus on product, market fit … If there is a token, it’ll naturally come.”

In a downshifted cycle, the venture message is clear: scale what works, experiment selectively and don’t confuse narrative with fundamentals.