Written by: CZ (Zhao Changpeng) Compiled by: Wu Shuo Blockchain

In this exclusive interview with Talking Tokens Podcast, Binance founder Changpeng Zhao (CZ) reviewed Binance’s journey from its founding in 2017 to becoming the world’s largest exchange five months after its establishment, and shared his experience in user- and community-driven platform development. CZ said that if he started over, he would have learned about law, compliance and politics earlier and would have launched a product faster. In addition, he also talked about the development space of legal currency deposits and withdrawals, stable coins, lending and RWA, and believed that encryption is not just an investment asset, but a basic technology as important as the Internet and AI. In the second half of the interview, CZ discussed AI Agent’s use of cryptocurrency to complete payments, the impact of AI on blockchain security and performance, and the eventual integration of traditional finance and on-chain finance.

The speeches made by the guests do not represent Wu Shuo’s views and do not constitute any investment advice. Please strictly abide by local laws and regulations. Audio transcription and translation are done by GPT and may contain errors.

CZ looks back at the opportunities and lessons learned in the early days of Binance

Jacquelyn: This year marks the 9th anniversary of Binance. Before founding Binance, you also worked at a large exchange. Looking back at 2017, the market environment at that time and your life situation at that time, what were the specific problems you wanted to solve? What did you feel other exchanges were not doing well enough at the time that ultimately led you to decide to start Binance yourself?

Changpeng Zhao (CZ): Before that, I had basically been engaged in order execution and exchange system-related work throughout my career. So when I looked at the exchange industry in 2017, I felt that the product could be better, the technology could be better, the matching engine could be much faster, the security could be much better, and the customer service could be much better. Exchanges at the time didn’t really prioritize users. So I think there's clearly a lot of room for improvement in this industry. However, we were lucky enough to catch the boom at that time and we seized the opportunity.

Jacquelyn: Do you think the CZ in 2017 and the CZ in 2026 are still the same person? If you could give your 2017 self one piece of advice right now, taking into account everything you already know today, what would you say to your then self?

CZ Changpeng Zhao: No, I have aged a lot. As for whether I have become wiser, I am not sure. But I did experience more challenges, so to speak. A lot has happened in the past 9 years, so I'm a little older now, both mentally and physically.

Based on what I know today, I would tell my younger self: Spend more time learning about law and compliance, but also learn about politics. Before 2017, I was just a technician. My focus is on building better products and protecting users. But I misjudged the importance of the legal aspect, which happens to be my weakness, so this is the piece of advice I would give myself. In addition, I didn’t know enough about international law at the time, and I didn’t know how different countries applied their own laws. Some U.S. laws apply globally, others do not, and the retroactive periods for these laws can be very long. So that's the first aspect I would remind myself of.

Secondly, I will tell myself to move faster. There were many things and many fields at that time, and I would hesitate whether to launch the product or not. In my experience today, this is not entirely hindsight. Just from an entrepreneur's perspective, a better approach is usually to bring the product to the market as early as possible and then continue to adjust based on market feedback. I think there are a few things we did relatively late. For example, Binance only launched futures business two years after its establishment. If I had to do it over again, I would probably launch a futures product earlier.

Topping the world’s No. 1 ranking in five months

Jacquelyn: This is indeed a good point. Many times, founders will stay in stealth development for a long time, hiding everything. When the product is officially launched on the market, they will expect an amazing effect like fireworks. But sometimes, the feedback from the market is: "We don't like this product," and it might be much better if they could launch the product earlier. However, Binance also surpassed dozens of exchanges at that time and quickly climbed to the number one position in the industry, right? Was it in year one or year two?

CZ Changpeng Zhao: In the first year, probably within the first five months of our establishment, we became the world’s largest exchange, and we have remained in that position every day since.

Jacquelyn: So to what would you attribute this success? Binance became No. 1 at that time, and has maintained this position for so many years, because it may be difficult to get to No. 1, and of course it cannot be said to be easy, but at least it is possible; but once it reaches the No. 1 position, it is actually very easy to fall, right?

CZ Changpeng Zhao: I think there are many different factors behind the success of any company. The product has to be good and the customer service has to be good. More importantly, you must have a mission. Our mission is to increase financial freedom, and one of our most important values ​​is protecting our users. These are very basic, very fundamental things. I think after being No. 1, protecting our users helped us continue to stay ahead, and it also drove our growth. We did get some luck in 2017 as most other exchanges were still primarily Bitcoin exchanges, with some US exchanges only listing 4 coins at the time. At that time, one of the largest exchanges in the United States had not even listed Ethereum.

So, we did take advantage of the opportunity, but it was not without competition. Bittrex was huge, Poloniex was huge, they were the two largest exchanges in the world at the time. So it’s not that there wasn’t competition at the time, it’s just that we caught the turning point in the industry and the momentum brought by altcoins. At the same time, our products are very good, our services are very good, and we take the protection of our users very seriously.

The wealth penetration rate of crypto assets is still less than 1%, and the industry is far from saturated

Jacquelyn: Totally agree. Compared with the past nine years, what do you think is needed to build a truly sticky product in today's market environment? In other words, what kind of product can make users come back again and again, instead of being like a revolving door where users come in and leave quickly?

CZ Zhao: I believe that no matter in any era or in any industry, there is always room for improvement in products, and nothing is perfect. For example, today, there is still very high friction in deposits and withdrawals of fiat currencies. If you can lower that friction in more parts of the world and lower the cost at the same time, I think that's one of the product directions that everyone is looking for. Looking at stablecoins, some of the most popular stablecoins do not provide users with income, that is, they do not pay interest to users. There are also stablecoins that pay interest to users, but making transactions with these stablecoins is difficult. Therefore, if there is a stable currency that can not only provide users with good interest income, but also allow people to freely use it for transactions, then this is another direction that can continue to be improved. RWA is also just getting started. Relatively speaking, it's still very new and there's still a lot that can be done. There are only a handful of stocks that are truly tokenized right now, and the overall market is still relatively U.S.-centric. What about the rest of the world? Which country now wouldn’t want to tokenize its stock market? Which country wouldn't want people all over the world to participate in its stock market, giving more than 300 million or more people access to investment opportunities? So I think there's still a lot that needs to be done. We're actually just getting started. A lot of people think the market is saturated, but looking at wealth penetration, I think the penetration of crypto assets is probably less than 1% and certainly less than 1%. So there is still a lot of room for growth in the future.

Jacquelyn: Speaking of sub-1%, we spent many years talking about "getting the next billion users into crypto," but in reality, even the first billion users haven't really entered yet. So what do you think is the biggest misconception in the industry right now about achieving mainstream adoption? What exactly does it take to truly achieve mainstream adoption?

Changpeng Zhao (CZ): I think that most people today, or at least a lot of people, still view cryptocurrency as an investment asset rather than an underlying technology. People buy Bitcoin and then discuss when to get out. But you wouldn’t say you’re going to “quit the Internet,” and you wouldn’t say you’re going to “quit AI.” You only use it because it's a technology. So I think that's a very common misconception. I see a lot of people looking at cryptocurrencies just as a speculative investment asset. They should look past this layer of appearance to see the underlying technology behind it.

Cryptocurrencies and blockchain are here to stay. This is a very basic technique. For me, it’s one of the three foundational technologies I’ve encountered in my life, the other two being the Internet and AI. So, the misconception that many people have is that they view crypto as a speculative investment opportunity and are constantly thinking about when to exit. But what really should be done is to take a long-term view of the project. I certainly admit that cryptocurrencies do have investment properties. But I always advise people to carefully look at the long-term prospects of this project before investing.

AI Agent will be paid in cryptocurrency

Jacquelyn: Speaking of AI, it seems to be becoming the next fundamental and transformative technology after encryption. We have seen this trend over the past few years. I'd love to hear your thoughts. Where do you think the two worlds of AI and crypto are really converging? Are there some intersections that most people haven’t discussed yet? Such as privacy, or other aspects.

Changpeng Zhao (CZ): There are several different points here. There are many different speculations about how AI and encryption will be combined and how they will work together. I can't guarantee that my judgment is correct, I'm just making some random guesses. But I think AI will eventually handle payments for us. Today, AI can help us search for the best hotels and the most suitable flights, but it can’t actually book it for us. I think it could happen within a few months, certainly not years. Once AI starts completing transactions for us, they will need a currency native to the technology.

Of course, they can also use your credit card to book hotels or flights online, but that would be very clunky and they should just use cryptocurrencies.

So I think payment may become the biggest breakthrough for AI. As more and more people use AI to complete payments, they will also use cryptocurrencies, either indirectly or directly. We can continue to observe what will happen in the end, and the traditional fiat currency payment system is not good at handling global payments. They may work well within your own country, but once cross-border payments are involved, problems become apparent.

I believe that there will be more and more global citizens in the future, and there will also be more and more globalized transaction needs. The problem may not be as obvious when people only trade within a country. But when you need to conduct a global transaction, such as paying someone on the other side of the world, or booking a flight to another part of the world, paying with cryptocurrencies is much easier. I think crypto payments may be the biggest breakthrough in the combination of AI and crypto. Of course, there are other directions of integration between AI and encryption. For example, AI can use blockchain for training data verification and so on. But I think these applications are still further away from real implementation. Most of the current AI companies are centralized companies. They are large, profitable, and highly valued, so they may not be particularly concerned about the decentralized nature of the technology right now.

Jacquelyn: Do you think crypto payments will appear in the near future? Or will it take another 3 to 5 years before it becomes part of the mainstream payments system?

CZ: I guess it may take another 1 to 2 years, at least that's the time I hope to see. Once people start paying with cryptocurrencies with the help of AI, they will find it much faster. Moreover, the first people to use such functions will most likely be users in the encryption industry. Crypto users are inherently early adopters, and AI is also a relatively new technology. Although many people are already using AI, the people who really start using AI to purchase goods and complete payments should still be mainly early adopters.

Enterprises should make full use of AI, but not rely too much on it

Jacquelyn: So, how do you see the impact of AI on Binance’s internal operations and infrastructure? What new products and use cases might it bring?

Changpeng Zhao (CZ): The way I look at AI is very similar to the way I look at the Internet. Can you say that you don’t want to use the Internet? This is probably not a wise choice. Maybe 20 years ago, a company could say that, but certainly not today. So I think basically any company, any individual, should use AI to the fullest extent possible within their capabilities. But at the same time, AI shouldn’t be overused. AI is still new and there are some things it’s not good at.

We should think of AI as a productivity tool. As long as it increases productivity, we should use it. I think AI is very good at helping people write code. It cannot yet complete a very large and complex system independently, but it can help people write code snippets, check for errors, and analyze code. So when it comes to programming, AI can play a big role. For those who know how to use AI, it can also do a lot in design, graphics, and video. Therefore, I think every company should use AI as fully as possible, but not overuse it. Some people will say: "Okay, I use AI for everything, all copywriting is written by AI, and all videos are generated by AI."

But I think we still live in a world composed of people, and people still need human participation and warmth. AI is already good at creativity, but it's not yet good at the kind of original creation that humans are capable of. So I think humans are still going to continue to play a very important role.

AI will make encryption more secure and will also promote digital currency to replace paper money

Jacquelyn: Speaking of which, at least for now, AI and creativity really aren’t fully integrated. Of course you can use AI to create a lot of exaggerated and crazy videos, but that’s probably not the main use of AI in the future, right? I'd also like to hear your thoughts further. In what ways will AI really significantly improve the crypto industry in the next 5 years? You just mentioned payments. In addition to this, are there any other directions that people are not discussing much at the moment? There may be some use cases that sound special, but they can actually significantly improve the operating efficiency of this industry.

Changpeng Zhao (CZ): I believe that AI will profoundly change the entire security landscape, especially in the field of technical security. AI is very good at finding vulnerabilities, and this ability can be used in both good and bad directions. Developers can use AI to find vulnerabilities in their own systems. In fact, I think this capability would make today's systems much more secure because it would help developers find vulnerabilities very quickly. From what I've heard, Anthropic's latest released model is pretty powerful. I hope that this kind of model will soon be available in some version for developers to use to improve their systems. In the short term, there may also be a few hacks here and there.

Also, from a productivity perspective, blockchain is still limited by efficiency and throughput. We need blockchains that are faster and capable of carrying higher capacities, and if this can be achieved, the cost of using blockchains will also fall. I think AI will play a very important role in accelerating the development of these types of technologies. Beyond that, I think AI overall will create a more digital world where we use AI very deeply. If the Internet has brought us into the digital world, then AI will advance this process 10 times or even 100 times. But the more digital the world becomes, the greater our need for digital currencies becomes.

We don’t need paper money, the concept of paper money is very outdated. While most of today’s currency no longer actually exists in paper form, the entire monetary system still follows concepts from the paper money era, and we need to move away from this model once and for all. I think AI will push us past this tipping point. We can continue to see what happens next.

Crypto platforms can still expand lending and RWA

Jacquelyn: What is a product or use case that Binance isn’t launching today that you think millions of people will be using by the time Binance reaches its 18th anniversary, which is in another 9 years?

Changpeng Zhao (CZ): I’m not sure I can imagine things that far ahead. And since Binance has so many different products now, it’s a really hard question to answer. I would say that Binance, or any crypto platform, still has a lot of new ground to get into. Such as loans and debt. Binance already offers related products to some extent, but loans and debt still represent a very small proportion of the overall crypto industry. In the traditional financial market, debt is an extremely large field, but it has not really entered the encryption industry yet. The more obvious direction may be the tokenization and RWA I mentioned earlier. Everyone is discussing these things now. As for something that no one has thought of or done, I really can’t think of it at the moment. I'm not very good at coming up with new ideas out of thin air that no one has thought of yet.

Relearning the crypto industry after stepping down as CEO

Jacquelyn: You have moved from being the founder and CEO of Binance to another role in the ecosystem. I want to know, after you retired, has the way you view this industry changed? Are there any opportunities or issues that you are aware of now that you might have overlooked when you were running Binance operations on a day-to-day basis?

CZ Changpeng Zhao: I think this does bring a very interesting and completely different perspective. When I was running Binance before, everything would become: "How does this impact Binance?" So the perspective was much narrower at that time. You’re in 20 to 25 meetings a day, questions are coming up, and everything is very busy. Binance is again a very large and influential platform, so everything else you do will often seem small in comparison. But now, as I've been forced to take a step back, I'm starting to look at things more from an industry perspective. Now I have time to learn new things, such as AI, biotechnology, and I can also gain a deeper understanding of some new directions within the crypto industry.

A lot of people have a misunderstanding: they think that when I was Binance CEO and responsible for company operations, I would be very knowledgeable about all the new things in the crypto industry. But in fact, I was so busy that I didn’t even have enough time to seriously learn new things emerging in the crypto industry. So now, I do have a broader perspective, which also allows me to learn new things faster. For example, when I was managing Binance, I actually missed stablecoins to some extent. I did not expect that stablecoins would develop to the scale they are today, but they did.

I originally thought that stablecoins were just a temporary patch technology, mainly used to connect and complete some transactions between different crypto exchanges, and that they might not develop much after that, but it turns out that I was wrong at the time. Now, I spend a lot of time working on RWA, which is the tokenization of assets, and I also advise some governments in this area, so I think I do have more time now to observe and study new things.

Jacquelyn: If you were to start from scratch today, without Binance and without its current reputation, but with only the experience accumulated in the past, what would you choose to build?

CZ Changpeng Zhao: Unfortunately, if I want to build a new project, I still have to choose something that I know how to do. And my experience basically only brings me back to the direction of building exchanges.

Jacquelyn: Got it, so it will still be an exchange.

Changpeng Zhao (CZ): Yes, I always tell others that people should find the intersection of three things: what they are good at, what they are interested in, and what can create value for others. But for me, what I do best, and most of my experience, comes from building trading systems. If I was asked to lead an AI team, it would probably be a disaster, so I still have to stick to doing what I know best.

Traditional finance and on-chain finance will eventually merge

Jacquelyn: In the long term, do you think traditional finance, on-chain finance, and these different transaction ecosystems will eventually merge into a whole?

Changpeng Zhao (CZ): Of course, I think there should be no opposition between so-called “crypto finance” and what we call “mainstream finance” or “traditional finance”. For example, you wouldn't call Internet communication a different communication system alone. There used to be paper mail, now there's email, but today, when we want to send a message to someone, most of us don't use traditional post anymore, so the two eventually need to be combined. There is no dichotomy between traditional finance and crypto-finance. Finance is finance. Cryptocurrency and blockchain are just a new technology adopted by the financial system. Because the technology is still so new, a separate industry segment has temporarily developed around it, but this industry segment will eventually be integrated into the entire financial system.

We are already starting to see this integration, stocks are being tokenized and moved to the chain; traditional banks and financial institutions are also using blockchain. So I think it's all going to come together eventually, that's for sure. There should not be two financial industries, nor should there be two parallel financial systems. In the end, there should be only one financial system.