对话付鹏：比特币不是数字黄金，它是AI资产
对话付鹏：比特币不是数字黄金，它是AI资产
By Lin Wanwan
On April 23, 2026, at the Hong Kong Convention and Exhibition Centre, economist Fu Peng (付鹏) delivered a keynote on the convergence between FICC (Fixed Income, Currencies, and Commodities) and crypto assets. It was also his first public dialogue since joining New Huo Group earlier this year as Chief Economist—news that coincided with a notable move in the company’s share price. Fu waved it off: the market, he argued, wasn’t “buying him,” it was beginning to price in a single idea—FICC + C, where C stands for Crypto.
One line from his talk quickly became the most repeatable: integration is one-way. In other words, crypto doesn’t “absorb” Wall Street; crypto is being forced to interface with Wall Street’s standards—custody, compliance, risk control, disclosures, and settlement discipline. That framing is more useful than it sounds, because it clarifies what the next phase of blockchain adoption is really about: not ideology, but infrastructure.
Even more provocative was his assertion that Bitcoin is not digital gold—it is an AI asset.
This article unpacks what that means, why FICC matters to crypto users, and how to think about Bitcoin, tokenization, and self-custody as the market shifts toward institutional rails.
1) Why “FICC + Crypto” is the real convergence
If you want to understand where the deepest liquidity and the strictest rules live, don’t start with spot equities or meme cycles—start with FICC.
FICC markets dominate global capital formation and macro hedging. They are where governments fund deficits, where rates get priced, where FX risk is transferred, and where commodities are financed. Crypto, by contrast, is still young: high-speed innovation, but historically uneven market structure.
So “FICC + C” is not a marketing slogan. It’s the industry’s attempt to answer a hard question:
Can crypto assets be integrated into the world’s most regulated, risk-managed financial stack without breaking either side?
To understand the traditional side of that stack, see an overview of the FICC market structure from a mainstream finance reference such as SIFMA’s FICC resources or a primer on fixed income markets.
2) “Fusion is only one-way”: what that implies for crypto builders and users
Fu Peng’s “one-way fusion” claim is less about surrender and more about constraints. Institutions do not adopt assets the way retail communities do. They adopt interfaces:
- Qualified custody and operational controls
- Clear asset classification and risk disclosures
- Auditable reserves and transparent issuance rules (especially for stablecoins)
- Market surveillance, best execution, and robust settlement
- Standardized collateral and margin frameworks
This is why so much of the “boring” work in crypto—custody tech, proof-of-reserves, on-chain analytics, compliance tooling, and tokenization standards—has become strategically important.
From a macro perspective, tokenization is the bridge. The Bank for International Settlements has repeatedly highlighted tokenization and unified ledgers as a direction for modern finance (see the BIS work on tokenisation and the future of money and settlement).
Key takeaway: the next wave of crypto adoption will not be driven only by narratives. It will be driven by compatibility with FICC-grade risk and settlement expectations.
3) Bitcoin is not “digital gold”? A more useful comparison for 2026
Calling Bitcoin “digital gold” helped early adoption: scarcity, neutrality, and a hedge-like posture. But it can also trap Bitcoin inside a single mental model—static, inert, primarily a store of value.
Fu Peng’s alternative framing—Bitcoin as an AI asset—is controversial, but it points to a real shift: markets are entering a phase where capital, liquidity, and even execution may be increasingly mediated by software agents.
Bitcoin’s properties are uniquely compatible with that world:
- Machine-verifiable scarcity: transparent issuance and a predictable supply schedule (see the original Bitcoin whitepaper).
- Always-on liquidity: global, 24/7 trading and settlement.
- Neutral collateral potential: not issued by a corporation or a state; easier to treat as “external” collateral in multi-party systems.
- Programmable integration at the edges: even if Bitcoin’s base layer is conservative, the financial system around it (custody, wrapped representations, derivatives, and settlement networks) can be engineered for automated workflows.
In an AI-driven market structure, the most valuable assets are not only the ones humans “believe in,” but the ones machines can price, pledge, hedge, and settle with minimal ambiguity.
So “AI asset” doesn’t have to mean “AI token.” It can mean:
An asset that becomes strategically important in a world where decisions are increasingly automated, portfolios are continuously rebalanced, and collateral moves at software speed.
4) What “AI asset” could mean in practice (without the hype)
The phrase can be abused, so it’s worth pinning down reasonable interpretations.
4.1 Bitcoin as high-grade collateral in automated finance
As prime brokerage, margin, and collateral management become more automated, the system tends to favor assets with:
- deep liquidity,
- simple rules,
- global accessibility,
- minimal issuer risk.
Bitcoin arguably fits that profile better than many alternatives—especially as institutional adoption progresses through regulated products and custody standards.
4.2 Bitcoin as a settlement anchor in a fragmented tokenized world
As more assets become tokenized—fund shares, treasuries, commodities claims, and money-market representations—markets may end up with many forms of “digital value.” In that scenario, neutral settlement assets matter.
This is not guaranteed, but it aligns with the broader tokenization direction discussed by global financial institutions (see the BIS link above) and real-world pilots like Hong Kong’s tokenized bond initiatives (for background, see the Hong Kong Monetary Authority’s information on tokenised green bonds).
4.3 Bitcoin as “reserve asset” for autonomous agents
A more speculative angle: if autonomous agents (AI traders, treasury bots, cross-border commerce software) need a reserve asset, they will prefer assets that are:
- censorship-resistant enough for global usage,
- liquid enough for fast conversion,
- and simple enough to audit.
Bitcoin is one of the few candidates that can plausibly meet those constraints at scale.
5) The uncomfortable part: AI also amplifies volatility and reflexivity
If AI increases market efficiency, it can also increase market speed. That has consequences:
- Faster narrative rotation: news-to-trade latency collapses.
- More systematic leverage: automated strategies can crowd into the same trades.
- Sharper liquidations: when risk models align, drawdowns can accelerate.
For retail users, the implication is not “avoid crypto.” It’s: treat risk management as a first-class feature.
Practical questions to ask yourself:
- If Bitcoin is used more like collateral, what happens during a liquidity crunch?
- If tokenized assets spread, what is the weakest link—issuers, custodians, bridges, or stablecoin reserves?
- If AI agents trade faster than humans, do you have a plan for position sizing, custody, and exit liquidity?
6) Where blockchain fits: tokenized assets, stablecoins, and on-chain settlement
To connect FICC and crypto, three layers matter most:
-
Tokenized assets (RWA tokenization)
Bringing real-world instruments on-chain is not just a narrative; it’s an operational change: issuance, transfer restrictions, corporate actions, and reporting can become software. The challenge is legal enforceability and trustworthy settlement. -
Stablecoins and tokenized cash
Stablecoins function as on-chain liquidity and are central to market plumbing. But they also concentrate risk in reserves, governance, and regulatory interpretation. Users should track issuer disclosures and redemption mechanics via primary sources such as The Financial Stability Board’s work on global stablecoin arrangements (reference jump). -
Custody and key management
As more value becomes tokenized, custody becomes the control plane. In a “FICC + C” world, the operational question—who can sign, under what policy, and with what auditability—matters as much as price.
7) The user-level playbook: how to participate without losing the plot
If you accept the “one-way fusion” thesis, the winning behavior for users is not to chase every new product. It is to build a resilient setup that survives both innovation and regulation.
A practical, non-exhaustive checklist:
- Separate trading from long-term holdings: keep long-term reserves in cold storage; keep only what you need hot.
- Treat stablecoins like credit instruments: diversify, monitor disclosures, understand redemption routes.
- Prefer transparent protocols and clear risk surfaces: complexity compounds quickly in AI-driven markets.
- Assume phishing and social engineering will get smarter: AI helps attackers too.
This is where Bitcoin custody and hardware-based self-custody become less of a hobby and more of a baseline security practice.
8) Why self-custody becomes more important when Bitcoin becomes “AI-native collateral”
If Bitcoin increasingly behaves like high-grade collateral in automated systems, then custody isn’t merely about “not your keys, not your coins.” It’s about controlling the one action that matters most in any on-chain financial system:
the signature.
A hardware wallet reduces exposure by keeping private keys offline and isolating signing from compromised computers or browsers. For users participating in DeFi, on-chain settlement, or long-term holding, this can be a meaningful upgrade in operational security.
OneKey, as a hardware wallet option, is designed around self-custody and multi-chain usage, with an emphasis on transparent security design and practical day-to-day signing workflows. If your thesis aligns with Fu Peng’s—where crypto integrates into more institutional-grade rails—then upgrading your personal custody stack is a rational step, not a maximalist statement.
Conclusion: “Digital gold” was the beginning, not the destination
Fu Peng’s talk in Hong Kong offered a useful lens: crypto is moving toward FICC-grade integration, and the integration is largely one-way—toward stricter interfaces, better custody, clearer settlement, and more disciplined risk frameworks.
His “Bitcoin is an AI asset” line is not a prediction that AI will magically pump the price. It’s a claim about fit: in a world where software agents allocate capital and collateral moves continuously, Bitcoin’s neutrality, scarcity, and liquidity make it unusually compatible with machine-mediated finance.
If that’s the direction of travel, the smartest move for most users is also the least flashy: understand the rails, respect the risks, and take custody seriously.
