In the crypto industry, there are actually very few people who have genuinely participated in building the foundational narratives and prototype rules. They appeared early, helped shape the direction, and personally witnessed the divergence of rules. As industry cycles shifted, many chose to fade into the background—occasionally mentioned by name, yet increasingly becoming legends from the industry’s early days. Han Feng stands apart: he has remained squarely in the arena.

After most others withdrew from the public eye, he stayed active within the industry, consistently focusing his attention on the hardest—and most fundamental—problems. In early 2026, he published his new book Stablecoins: Exploring the Future of AI Agent Economies, and was concurrently elected as a Corresponding Member of the U.S. National Academy of AI—alongside Jensen Huang.

This dual recognition reflects the core question he has long pursued: How is value anchored? And how can a system sustainably operate over the long term without a centralized arbiter? This conversation begins with today’s most concrete challenges—the stablecoin dilemma, Hong Kong’s market choices, and the intersection of AI and Web3—before ultimately pointing toward a more fundamental question: Are we building a truly new order—one capable of supporting future AI agent economic activity and fundamentally distinct from the fragility of traditional finance?

Han Feng anchors the discussion in stablecoins—the critical infrastructure bridging the crypto world and the real economy. In his view, current mainstream stablecoin approaches, regardless of technical form, have fundamentally failed to fulfill the role of “future economic value benchmark.” The fiat-collateralized model exemplified by USDT still rests trust on the issuer’s credibility: whether assets truly exist, how they’re custodied, and how risks are managed—all remain unverifiable black boxes. “This isn’t a new financial order,” Han Feng points out, “it’s merely transplanting traditional finance’s credit structure onto-chain.”

The alternative path—such as MakerDAO’s on-chain over-collateralized model—advances further in decentralization, yet relies on smart-contract-triggered “forced liquidations” to maintain stability. Under extreme volatility, this mechanism not only harms individual users but may also amplify systemic risk. “It applies deterministic code rules to inherently uncertain, complex economic behavior—a structural lack of resilience.”

A key contradiction thus emerges: a system striving to build a decentralized new economic order relies, at its most critical value-transmission unit (the stablecoin), either on the centralized credit of the old world—or on automated rules lacking judgmental capacity. Han Feng describes this as conducting global settlement in the digital age using the logic of paper ledgers—a fundamental architectural rift between vision and implementation.

So where should the true “anchor” be placed? His answer points directly to consensus itself. “If we seek a truly global, censorship-resistant, decentralized trust foundation, the Bitcoin network is currently the only answer rigorously tested over time.” To him, Bitcoin’s significance lies not in its market cap or historical stature—but in its mode of operation: maintained collectively by a global community, requiring no centralized arbiter; governed by highly predictable rules, with an issuance path immune to human intervention. It is precisely this long-term self-constraining architecture that has, across multiple extreme market cycles, gradually revealed a rare kind of stability.

If Bitcoin solves “where trust comes from,” the next question is: how does the system maintain stable value over the long term amid inevitable price volatility? To address this, Han Feng proposes AiFi (AI-Driven Finance). Its goal is not to amplify returns—but to build a resilient value-operating system. Within this framework, system operation moves beyond passive rule execution by smart contracts; instead, AI agents actively manage volatility: dynamically hedging before risks escalate; adjusting asset allocations across varying liquidity environments; and continuously fine-tuning system parameters based on multidimensional signals—to preserve overall stability.

“AI doesn’t provide definitive answers—it provides space for judgment,” Han Feng notes. “For the first time, it allows the system to operate like an economy, rather than a static set of rules.” In his vision, AI does not replace decentralized consensus—it serves as a regulatory layer on top of consensus, enabling a Bitcoin-based system to function sustainably over the long term. This idea is now undergoing real-world validation through BTCD, a Bitcoin-native stablecoin system he is developing.

“Any sufficiently radical system design must ultimately confront reality.” In Han Feng’s view, the rift between the crypto world and the real world stems not from technological insufficiency—but from the “gravity of reality”: the convergence of legal consensus, institutional enforcement, and user habits. He stresses that the real challenge isn’t whether a system is decentralized—but whether technical consensus is permitted, usable, and routinized in daily life. That is the meaning of the “last mile”—not inclusion in a whitepaper, but integration into lived experience.

Accordingly, Han Feng exercises extraordinary restraint in choosing practical implementation points. By contrast, Hong Kong presents a rare positional advantage: it bridges mainland China’s capital and demand with Southeast Asia and global markets; operates under a mature legal framework, yet retains clear, experiment-friendly regulatory space. “It’s not the destination—it’s the interface.” To Han Feng, Hong Kong functions more like a connector between institutions and technology—a testbed for consensus落地—where new systems are, for the first time, permitted to undergo real-world scrutiny.

Beyond building systems designed for longevity, Han Feng repeatedly emphasizes a question often overlooked: What role should the individual occupy within such a system? In his assessment, the central tension today lies not in bull-bear cycles—but in the rapidly escalating system complexity, while most participation methods remain stuck at an early stage. As infrastructure matures and institutions steadily enter, the room for gaining advantage via information asymmetry, narrative speed, or short-term sentiment is being structurally compressed. “The era of simple token launches and quick arbitrage is naturally exiting the historical stage.”

Within this environment, options for ordinary participants are narrowing. Thus, his advice—though not novel—is growing ever rarer: adhere to value investing; embrace long-termism. Han Feng states, “Time is the only dimension available to ordinary people to hedge against information asymmetry and capital advantages.”

While the industry relentlessly generates new narratives, Han Feng chooses to return to the most foundational—and most difficult—question: In the absence of a centralized arbiter, how is value defined, anchored, and sustained in circulation? The dialogue with Han Feng traces an exceptionally clear path: decentralization solves “where trust comes from”; AI solves “how the system runs long enough.” This is neither slogan nor rhetoric—nor does it require repetition. The answer is already written into the system itself.

[Klickl]