Author | MetaEra
On August 27, at the “Hong Kong CryptoFi Forum,” Changpeng Zhao (CZ), founder of Binance, the world’s largest digital asset exchange, laid out his forward-looking thoughts on the industry’s future development.
CZ focused on five themes: the evolution of stablecoins and the U.S. dollar’s strategic position; the regulatory and liquidity bottlenecks of RWA; the potential of decentralized exchanges; how the Digital Asset Treasury (DAT) model offers a new investment avenue for traditional investors; and how the fusion of AI and Web3 will transform trading paradigms.CZ’s views not only reflect his deep insight into current industry developments, but also reveal his strategic thinking about the future landscape of digital finance. These perspectives are important references for understanding trends in crypto finance and identifying investment opportunities.
What follows organizes CZ’s on-site viewpoints into an article; the author has sought to preserve CZ’s original phrasing as much as possible.
CZ on Stablecoins: From a Volatility “Safe Harbor” to a Tool for U.S. Dollar Globalization
To be honest, I’m not a specialist in the stablecoin vertical, but Binance processes roughly 70% of the world’s stablecoin trading volume, which makes us the most important distribution channel for stablecoins in the industry.
Let me briefly introduce the history of stablecoins. The earliest technical prototype was “Colored Coins,” the Bitcoin community’s first exploration of “on-chain assets.” In 2014, USDT was initiated by Brock Pierce. The project grew slowly at first; Pierce gradually stepped back, ceding to the current USDT team—Craig Sellars and others—and even as late as 2017, it still hadn’t taken off.
When Binance launched in 2017, we focused on crypto-to-crypto trading, supporting pairs like BTC/ETH and BTC/BNB, but without fiat on-ramps. That created a user-experience problem: whenever the bitcoin price fell, users had to withdraw BTC to other fiat exchanges to convert into fiat, and it was highly uncertain whether those funds would flow back to our platform.
This was also very unfriendly for users. To improve UX, we decided to support USDT as a “safe harbor” when markets were down. At the time, we viewed stablecoins as a short-term store-of-value tool, so the decision to support USDT was relatively simple—we didn’t sign complex partnership agreements or strategic deals; we just integrated the product.
From there, USDT entered a period of rapid growth:
First, after 2017, crypto-to-crypto exchanges grew quickly. Many platforms, including Binance, supported USDT, which propelled its expansion.
Then came a second growth wave: many users in Asia needed U.S. dollars but found it difficult to open USD accounts directly; USDT provided an alternative. Tether has always been highly profitable. Due to U.S. regulatory pressure and banking frictions, they’ve remained relatively low-profile.
In 2019, U.S. compliance firm Paxos proactively approached us to co-issue a stablecoin, which became BUSD. From 2019 to 2023, BUSD’s market cap grew to $23 billion. During this period, we invested relatively limited resources—mostly brand support and promotions, like “free withdrawals.”
In 2023, the U.S. government wound down the BUSD project. Had BUSD continued, it would likely have scaled well, because at that time BUSD’s growth rate was outpacing both USDT and USDC. It’s worth emphasizing that when BUSD was shut down, all user funds were fully redeemed—clear proof that BUSD was compliant, transparent, and safe.
Stablecoins and exchanges have become two of the most core profit centers in crypto finance. The business model is highly streamlined: once licensed, users deposit funds and the platform issues tokens; when users redeem, the platform returns cash. The model has low barriers, high liquidity, and huge market potential, with strong long-term profitability.
At the national-strategy level, the U.S. government’s attitude toward stablecoins has changed significantly in recent years. The current administration is very savvy—given its business background, it deeply understands Tether’s strategic value to the dollar’s global standing. Around $100+ billion in USDT funds are used to buy U.S. Treasuries, and Tether is widely used globally. The key point is that Americans themselves don’t need stablecoins—they can just use the banking ACH system for USD transactions. Almost all USDT users are outside the U.S., which actually expands the dollar’s global influence.
This aligns closely with China’s desire to expand the RMB’s international footprint. In essence, stablecoins help the base currency globalize, which should be very attractive to governments. Of course, as freely circulating blockchain assets, stablecoins do pose challenges to capital controls, but such issues can be addressed. Many of the dozen or so countries I’ve interacted with are very interested in developing local stablecoins—everyone wants to put their fiat currency on-chain.
When the U.S. passed the “GENIUS Act” in July and signaled constraints on the development of CBDCs, that reflected a far-reaching strategic plan to protect the dollar’s global primacy. Stablecoins are popular precisely because they circulate freely and offer great UX, whereas some government-issued digital currencies may be more stringent on surveillance and controls—reducing market acceptance. In fact, since 2014, more than 20 countries have tried to issue CBDCs, but none has truly succeeded at a market level.
Blockchain is fundamentally ledger technology, and its first application scenario is finance, so stablecoins are a natural fit. So far, only dollar stablecoins have matured; stablecoins for other currencies haven’t emerged, which means massive growth potential ahead. Now, every country wants to develop stablecoin businesses. I think each country should have at least a few stablecoin products.
CZ on RWA: The Triple Challenge of Liquidity, Regulation, and Mechanism Design
Although Real-World Asset (RWA) tokenization has vast market potential, implementation is far harder than the market expects. The challenges can be summarized in three areas:
1) The Liquidity Dilemma
Practically speaking, highly financialized products are easier to tokenize, because traditional financial products inherently trade easily and have mature digital representations. Non-financial assets face foundational obstacles. In theory, you can “tokenize everything”—cities, buildings, even individuals could issue tokens—but in practice it’s riddled with problems.
Take real estate. Even in Hong Kong’s relatively volatile property market, volatility is still small versus bitcoin. Once you tokenize a low-vol asset, its trading appeal is weak; order books lack depth. Liquidity drops, investors place fewer orders, and a vicious cycle follows: shallow books lead to low volumes. If an investor wants to move in or out with, say, over 100 million in funds, it’s nearly impossible to fill. Even if the asset is on-chain, liquidity is insufficient; you can trigger unexpected volatility—or even short-term manipulation—more easily.
2) Regulatory Complexity
Financial products raise a core question—are they securities? Are they securities, commodities, or something else?
In large or financially advanced jurisdictions, definitions and regulators are distinct; in some small countries, a single regulator may “do it all.” If multiple agencies are involved, compliance becomes complex. Companies may need multiple licenses: futures, spot, digital asset, bank custody, etc. The more licenses you have, the more constrained your business model is—and often you can’t even get a single line of business to work.
3) Product Mechanism Flaws
In my view, U.S. securities tokenization products are not yet sound at the product level. Consider tokenized stock products like xStocks: token prices don’t track real stock prices, which is unreasonable. In theory, if a price gap exists, arbitrage should close it. In reality, the spread persists—evidence the mechanism doesn’t work. In other words, in today’s tokenized-stock track, tokens and stocks aren’t truly linked, so the model isn’t yet viable at a product level. Although multiple tokenization approaches are being tried in the U.S., a genuinely workable solution hasn’t been found.
Despite these challenges, one RWA model clearly works—stablecoins. Their underlying assets are mainly U.S. Treasuries and other traditional instruments, and their success validates the feasibility of tokenizing financial assets.
The dollar is already on-chain through stablecoins. In today’s blockchain ecosystem, almost everything is USD-denominated; the euro and RMB are essentially absent. As the world’s largest equity market, the U.S. could use blockchain to attract global investors to U.S. stocks, which would be hugely beneficial. If U.S. equities can be put on-chain, America’s leadership in global financial markets would be further reinforced.
Rationally, the U.S. should embrace this; other countries that don’t participate risk marginalization. For example, the Hong Kong Stock Exchange—an exchange of global significance—could see its influence erode if it sits out this wave. The Shanghai Stock Exchange and other Asian bourses face similar strategic choices.
Economically, this is 100% the right thing to do—refusal means elimination. Just as, absent Alibaba, China’s e-commerce might have been fully dominated by Amazon, failing to act in fintech could bring profound economic consequences.
Regulatory challenges aside, the economic impact is so deep that every country should seriously plan. With Asian ingenuity and innovation, these issues will be solved; timing will be key.
For businesses and entrepreneurs, you must hit the market window precisely: too early and survival is hard; too late and the opportunity is gone.
Right now is a rare golden window. U.S. policy is showing unprecedented support for crypto, which will inevitably spur other growth-oriented countries to move. As Asia’s long-time financial hub—and with a supportive government—Hong Kong has a rare historical opportunity. Everyone should seize it.
Exchange Transformation: Decentralization Will Surpass Centralization—How Should Hong Kong Seize the Chance?
1) The Essence of an Exchange and a Vision for the Future
I believe exchanges shouldn’t limit what can be traded; all assets should circulate freely on a single platform.
Once assets are on-chain, they’re just tokens—whether crypto-native or RWA, there’s no substantive difference from an exchange technology perspective. Adding a new asset class usually doesn’t require heavy development—just support the existing chain. Most RWAs don’t need their own blockchains; they issue on Ethereum, BNB Chain, Solana, etc. So wallet and exchange support is straightforward. The real delta lies in compliance: which regulator licenses you, and whether approval is granted. Once licensing is solved, there are virtually no technical hurdles.
In the long run, future exchanges should enable unified trading across global asset types. A building, a celebrity’s future IP revenue, even a person’s net worth—everything should trade in one market. That maximizes liquidity and makes price discovery more efficient.
That said, RWAs do have unique hurdles. If you tokenize a building, and later want to sell it, you may only be able to sell part of it. Once tokens are issued, if even one holder with a single unit refuses to sell, you can’t fully buy back the building—or the cost becomes huge. Think of it as an “on-chain holdout.”
While “global asset on-chain” will take time, for ~90% of countries it’s not out of reach. Compared to very complex regulatory systems in some major countries, many nations may directly adopt unified international standards and be first to push global on-chain asset circulation.
2) Building a World-Class Exchange in Hong Kong
On how Hong Kong can build a world-class exchange, here’s the logic. Early on, many jurisdictions manage crypto risk with strict controls. Regulators fear mistakes, so they require everything local: local license, office, headcount, compliance, servers, data storage, matching engine, user base, even wallet infra—all isolated from overseas.
In the physical world, that’s easier—vaults and air-gapping. In digital assets, it doesn’t matter much. Whether servers are in Hong Kong, Singapore, or the U.S., the likelihood of hacker attacks is the same—everything runs online.
More importantly, splitting operations means just the wallet infrastructure alone can cost on the order of $1 billion to build securely. The real bottleneck isn’t money but talent—you can’t easily hire hundreds of world-class security experts to replicate the stack. Duplicating a full system essentially costs as much as building a top-tier global exchange.
From a liquidity standpoint, if you only allow locals to trade—for Hong Kong’s 8 million people, or a small country with 200–300 thousand active users—you won’t generate sufficient volume. Without liquidity, price swings are extreme, which harms users.
Real user protection comes from deep order books—so a nine-figure order doesn’t blow through price, and derivatives don’t auto-liquidate in volatile moments due to market depth. On a low-liquidity venue, buying 10 BTC incurs high slippage and cost. Large, global exchanges provide baseline user protection—lowering trading costs.
When countries try to build isolated systems, management complexity explodes, making little business sense. Meanwhile, many jurisdictions restrict tradable assets; for example, Hong Kong currently has many listing constraints and limited product coverage. As far as I know, most licensed exchanges in Hong Kong are loss-making; while they can sustain in the short term, that’s not viable long term.
However, Hong Kong has an advantage—speed of iteration. We saw a new stablecoin draft in May, even earlier than the U.S. Government is very proactive in engaging industry participants, including conversations with people like us. Hong Kong may have been conservative in previous years—understandably—but as the global landscape shifts, it’s now very proactive.
I think this is a great starting point. Past constraints don’t predetermine the future; now is a prime time to explore. That’s why many Web3 builders, myself included, are looking for opportunities in Hong Kong.
The Future of Decentralized Exchanges
I believe decentralized exchanges (DEXs) will inevitably outscale centralized exchanges (CEXs). While Binance may be large today, I don’t think it will remain the largest forever.
DEXs currently don’t require KYC. For users who can handle wallets, they’re very convenient and fast, and highly transparent—sometimes too transparent, since everyone can see each other’s orders.
● Regulatory: We paid a heavy price because our KYC at the CEX level wasn’t good enough. At the same time, the U.S. doesn’t appear to have many measures targeting DeFi right now, which could bring a regulatory dividend for DeFi. Given my history, however, it’s hard for me personally to re-enter that space.
● User Experience: DEX UX is decent, but users must know how to use wallets. Those who have used CEXs know the UX there wasn’t ideal either—interfaces full of addresses, contracts, and “gibberish,” frequent block explorer checks, and defenses against MEV, etc. I faced many attacks while learning myself.
So for users moving from Web2 to Web3, most will start with CEXs—email-plus-password logins and customer support feel familiar. Over time, as more users become comfortable with wallets, some will shift to DEXs. DEX fees are actually higher today, but long term—thanks to technology—DEX fees should become cheaper.
Many DEXs now use token-based incentives. But these will end—no one can keep inflating forever, because unlimited issuance depresses price.
So we are still relatively early; token incentives exist. In the long run—say 5–10 years—DEXs will become very large. In 10–20 years, I believe DEXs will definitely surpass CEXs. That’s the trajectory.
Although I won’t lead such projects now, from an investment standpoint, we have invested in many similar efforts, usually with small stakes—we support from behind. I think this space has substantial room to grow.
CZ on DAT (Digital Asset Treasury) Strategies: A Bridge for Traditional Investors Entering Crypto
Many people oversimplify DAT (Digital Asset Treasury) as a concept, but in fact this track is quite detailed. Ultimately, its core logic is to package crypto in an equity-like form, allowing traditional stock investors to participate more easily.
There are multiple layers and forms in DAT, much like traditional companies—different models can coexist. Crypto ETFs are mainly issued in the U.S., but many investors lack U.S. brokerage accounts or are unwilling to bear high trading and management costs. By contrast, publicly listed companies like Strategy can allocate assets at lower cost by directly holding crypto. Their financing routes are more diverse as well—they can raise in the U.S., Hong Kong, Japan, etc. Differences in financing channels and investor bases by region shape unique market structures.
Within the listed-company model, DAT companies generally operate in the following ways:
1. Passive single-asset holding
Represented by Strategy, focusing on passively holding one asset—bitcoin. The approach is simple, with low management and decision costs; they can stick to the strategy through price cycles.
2. Active single-asset trading
Still one asset, but the management approach differs radically. These companies try to trade actively, making directional calls. You must evaluate the manager’s trading ability; with subjectivity involved, outcomes can be positive or negative.
3. Multi-asset portfolio management
More complex DATs hold multiple coins. Managers must decide allocation—how much BTC, BNB, ETH, etc.—and how often to rebalance. This tests managerial skill.
4. Ecosystem investment and construction
The most complex model: beyond holding coins, they deploy 10%, 20% or more into ecosystem investments. An ETH-focused company, for example, might invest to help the ecosystem grow. BNB and other ecosystem projects do similarly, but this demands stronger management.So DAT is not “just hodling.” Different models carry different management costs and requirements.
The DATs we currently support lean toward the simplest, first type. We prefer companies focused on a single asset—especially BNB—because evaluation is straightforward and day-to-day involvement is minimal. In bull markets, listed companies generally benefit; in bear markets, especially in the U.S., litigation risk rises. If the strategy is clear and simple, litigation risk is lower and legal costs fall—lawsuits are very expensive.
Our goal is to minimize operating costs and promote long-term holding. We don’t want companies to allocate into side investments; we hope they engage more deeply by supporting ecosystem development.
The significance of the DAT model is that many corporate finance departments, public companies, and even SOEs or central enterprises cannot directly buy crypto. Through DATs, we can effectively provide them with crypto exposure. This is a huge market—much larger than the native crypto crowd.
In DAT projects we’re involved with, we usually play the role of a small supporter. Most of the capital comes from traditional equity markets or other channels, which strongly supports our ecosystem by bringing in outside buyers.
We generally don’t lead or manage these companies. Instead, we leverage our network to find suitable managers. Running a public company is not our specialty, but many in the industry have that experience; we prefer to collaborate and create synergies.
AI × Web3: The Path from Concept to Reality
Frankly, AI × Web3 is not ideal yet. But I believe this trend is not mere hype—it will see breakthrough developments. A few months ago, I posed a question: what currency will AI use? The answer obviously isn’t the U.S. dollar or traditional rails, because AI can’t complete KYC. AI’s monetary system must be crypto and blockchain-based—paying via API calls or broadcasting transactions.
That implies exponential growth in blockchain transaction volume. In the future, each person may have hundreds or thousands of AI agents running in the background—video production, multilingual translation, content distribution, bookings, message replies, etc. Their frequent interactions will create massive micro-payments; crypto transaction volumes could conservatively grow by thousands of times. For example, a blogger can make the first third of an article free and charge ¥0.1 per read for the remaining two-thirds. If hundreds of thousands pay, that’s tens of thousands in income—impossible in traditional finance but easy with AI × Web3.
Transactions will also be more global. I could simultaneously hire engineers and designers from China, India, and around the world; AI will handle settlement and payment automatically. However, most so-called “AI agents” in Web3 today are still memecoin-style pseudo-products: glossy front-ends that simply call a mature large-model API like ChatGPT on the back-end, lacking real utility. What we need are AI tools that actually get work done and create economic value—and the top model companies are exploring this.
But AI requires enormous capital. The compute race for large models is brutal and costly. Reportedly, OpenAI has about 1–2 PB of compute; each PB costs around $6.5 billion per year, and they plan to expand by 10× to 100×—astronomical sums, not including chips. No VC, company, or even country can shoulder that alone. That’s why the AI industry is exploring new financing paths through Web3.
Fundamentally, AI should be treated as a public good. Many current large models are too closed. Letting token holders share in the upside—making models more open, decentralized, and broadly owned—may be a more reasonable direction. I’ve discussed this with several top model founders. While it’s still early, the trend will come.
Even though AI × Web3 is not yet mature, its future prospects remain highly promising.
Follow us
Twitter: https://twitter.com/WuBlockchain
Telegram: https://t.me/wublockchainenglish
best commentary on Global blockchain macro-economics with deep insights into Asian region developments