Written by: OKX

This article is an exclusive interview with Amber Group CEO Michael Wu, part of the OKX interview series "Friends of OKX." Michael reflects on the development journey of Amber Group, including its rapid expansion in 2021 and the survival challenges and profound reflections faced during the collapse of FTX and Three Arrows Capital in 2022. He shares insights on the team's transition from traditional financial quantitative trading to going all-in on Crypto, analyzes the essential logic behind market makers providing liquidity, and emphasizes the importance of compliance in mitigating long-tail risks. Furthermore, he prospectively discusses how the decentralized and permissionless characteristics of blockchain will provide a foundational pricing and payment network for the future AI Agent-led economy. Lastly, Michael Wu shares his experience as an entrepreneur dealing with pressure and the core criteria for investing in incubated projects.

The views expressed by the guest do not represent Michael Wu's opinions and do not constitute any investment advice. Please strictly adhere to local laws and regulations.

From Morgan Stanley to Crypto Builder

Mercy: Michael Wu, it's a pleasure to have the opportunity to interview you today. Do you speak Chinese or English more often?

Michael Wu: Both. I currently live in Singapore. I went abroad to study at 15 and completed my undergraduate studies in the United States.

Mercy: So, did you go directly into TradFi after finishing your studies in the U.S.?

Michael Wu: Yes, I went to Morgan Stanley, right on the 37th floor of this building. Our office is now above them.

Mercy: So you started your business near your old workplace?

Michael Wu: Not exactly, I was just moving offices and coincidentally found a suitable place here. It wasn't deliberate, just a coincidence, because I started working after graduating from university at the ICC (International Commerce Centre) in Hong Kong, which is this building.

Amber's Turning Point

Mercy: Actually, when we just talked, you mentioned that you have been an entrepreneur for nine years. When the Three Arrows Capital event happened in 2022, how many years into your entrepreneurship were you?

Michael Wu: About the fifth year.

Mercy: So such an event happening in your fifth year must have had a significant impact on you.

Michael Wu: Three Arrows had an indirect impact on us; the biggest impact was the FTX collapse at the end of the year. We started our entrepreneurial journey around September 2017, and Amber was my first entrepreneurial venture. We were lucky that we started preparing at the end of 2017, officially launched in the first quarter of 2018, and began to achieve profitability in the second quarter.

2021 was our fastest growth period. At that time, the whole industry was experiencing a boom, and we captured some of the sector's benefits. It was the first time the industry underwent true institutionalization, with many of our clients and investors beginning to treat Bitcoin and Crypto as a long-term investment asset. By 2021, we experienced explosive growth, with the company reaching a peak of about 1,300 employees globally, and we completed significant financing, including bringing in Temasek and the Brunei sovereign fund at the end of 2021, reaching a valuation of 3 billion USD.

Entering 2022, the market transitioned from a bull to a bear cycle. As you mentioned, the events of Three Arrows and FTX had substantial impacts on our company and the industry, which I believe were historic events in the sector. The impact on us was significant; in the first half of the year, I remember we were preparing for the entire company's Pre-IPO. At that time, two major investment banks were helping us with Pre-IPO financing, and we received our first Term Sheet for 8 billion USD. I told them that I believed we were at least a 10 billion USD company. I told the investment banks that I wanted 10 billion USD. About a week later, various liquidity crises began to occur in the industry.

We went from being a company that could potentially reach a 10 billion USD valuation if everything went smoothly to one where the entire market had dried up for financing. I remember that by the first quarter of 2022, exactly five years later, we returned to a state of loss for the first time in the second quarter of 2022. Then, in the second half of 2022, the market continued to deteriorate, directly impacting us as the FTX collapse resulted in significant financial losses for us. We were forced to undertake large-scale retrenchments, cutting back on business and scaling down. We also experienced a full year of the bear market in 2023, slowly starting to emerge from it and getting back on track. I believe that for me as an entrepreneur and for Amber as a company, in the long run, these experiences are valuable lessons.

Dual Test of Life and Entrepreneurship

Mercy: At the moment of that incident, what was your first feeling? Because I feel like your life has been quite smooth before this, right? Was that your first significant setback?

Michael Wu: You could say that. I think my co-founder has a similar background; we've both had relatively smooth lives with family, education, and career after graduation without facing significant setbacks. So, the collapse of FTX at the end of 2022 was probably the biggest setback in my entrepreneurial history and even in my life. Looking back, many things are quite incredible.

Aside from the FTX collapse, that year, my co-founder suddenly left at the end of the year, which was a massive shock to me. From the end of 2022 to early 2023, my focus was on resolving the company's difficulties and getting it back on track, which was manageable. But I remember that by the end of 2023, once everything was mostly settled, I found myself in a rather despondent state.

Everyone has a linear expectation of life, thinking it should continually rise, even accelerating. But encountering a significant setback for the first time, needing to get back up after falling, was a new experience for me personally and, I believe, for our still-young company.

Let the Company Survive First

Mercy: What do you think was the toughest decision you made at that time, which you now look back on as the right one?

Michael Wu: It's hard to say it was the best decision, but there were many tough decisions at that time. For example, scaling down the business, and we were forced to say goodbye to many colleagues. My primary factor was that the company had to survive, had to continue, and needed to keep moving forward.

One of the correct things we insisted on was: What kind of company are we? I always believed we were an ambitious and technology-driven company, always eager to be at the forefront of technology.

Decisions in Prosperity and Decisiveness in Adversity

Mercy: You must have faced enormous pressure then; how did you stay calm while making decisions?

Michael Wu: I think as an entrepreneur, those times are generally when you become calmer and find it easier to make decisions. Because when in difficult situations, your choices are limited. After careful consideration, you must weigh the pros and cons, and the decisions and executions you make become very urgent.

In contrast, during rapid industry growth or bull markets, there are often too many choices. Those times can be trickier because it requires more reflection; it seems like there are numerous good options, but truly great choices might actually be limited.

Returning to your question, I found it challenging during times like late 2022 when there were not many feasible options. Choosing the best from the limited options and executing them optimally is quite simple in that situation. It requires money, retaining essential personnel, and preserving core businesses and customers.

Mercy: Some of this thinking resembles our boss's, particularly your mention of navigating through bull and bear markets, as well as being product and technology-oriented; he thinks that way too.

Michael Wu: Yes, I have great respect for Star Xu; I think he is not only an OG in our industry but also a highly ambitious entrepreneur with a strong product vision.

From Quant Trader to Crypto Builder

Mercy: What was it about Crypto that attracted you to decide to leave such a glamorous job and go all-in on Web3?

Michael Wu: I feel a bit ashamed to say this, but I didn't leave Morgan Stanley to start a business because of Crypto. When we first started, we weren't in the fancy ICC building; we were in an old office with only two people, and below us was a seafood restaurant. At that time, we were called Amber AI.

Our entrepreneurial opportunity arose from my time as an FX and Rates Trader at Morgan Stanley. That was my first job after graduating, and in the first year or two at an investment bank, your main roles were to buy breakfast and coffee for everyone; you weren't allowed to take notes and had to remember what everyone on the desk wanted. I would sit next to more experienced traders to see how they conducted their trades. But I remember in about my second year, I was assigned my first real task, which was to manage Morgan Stanley's Asia EFX Book, which is their Asian currency electronic trading portfolio, responsible for programmed pricing.

Because I have a programming background and started coding at ten years old, my boss entrusted me with that task. This was significant because most pricing was still done through Voice Trading, which is a very traditional approach where traders quote manually or verbally.

At that time, the reason I was entrusted with this was that Asian currencies weren't mainstream and the portfolio was small. From 2014 to when I left Morgan Stanley in 2017, in just four years, the electronic trading volume surpassed all traditional quoting.

During those years, the Internet was booming, with the rise of electronic payments and digital banks, while the entire traditional financial industry was developing at a very slow pace. Particularly in 2017, when Google DeepMind's AlphaGo played Go, this AI had a huge impact on me. I realized that technologies like Machine Learning and AI would inevitably revolutionize the entire finance industry.

Therefore, we ventured into entrepreneurship in 2017 intending to do just that. We aimed to be among the earliest Asian companies using Machine Learning and Neural Networks to conduct pricing, liquidity management, risk control, and trading. Many large companies were reluctant to use such algorithms as they were termed "black boxes," making the processes unexplainable and thus they were unwilling to take liability, whereas, as a startup, we were willing to pursue such innovations.

So, how did we enter Crypto? In the same year 2017, when we started our business, we initially handled various assets, but two or three months later, we discovered that 95% of our revenue and profit came from a single asset: Crypto. I remember around the Lunar New Year of 2018, I discussed with my co-founders for several days, and there were only about ten of us at that time, and we finally concluded: All in Crypto.

Coincidentally, in the first quarter of 2018, the market had already transitioned from bull to bear, with Bitcoin dropping by around 50% and other Altcoins, including Ethereum, dropping by 80-90%. However, we primarily focused on a market-neutral asset management product, where the market was still very inefficient, presenting ample arbitrage opportunities.

I remember on my birthday in 2018, Bloomberg Tech had a headline article stating, "While the market crashes, some institutions can still achieve significant profits," referring to us.

Our brand immediately gained traction. Thanks to our background, we already knew many family offices and former investment bankers. They approached us wanting to invest in Bitcoin and similar Crypto assets, and we leveraged that wave of interest to establish our business targeting institutions, high-net-worth individuals, and family offices, akin to traditional investment bank Private Banking. This remains our main business today, and our NASDAQ-listed company, WhaleFin (originally listed as Emerald), also started from this.

We were indeed quite lucky in our early years; often, it was a case of right time, right place, right people. Interestingly, I got exposed to Bitcoin back in 2015 due to work reasons. When I was a trader at Morgan Stanley, my boss, who had no understanding of it, asked me to research it. I discovered it was genuinely feasible when Bitcoin was around 1,000 USD.

Reflections on Rapid Expansion and Long-Termism

Mercy: From that experience, what do you consider your greatest growth? Is there something you now realize was influenced by that experience in your decision-making?

Michael Wu: I definitely can come up with several points; one of the most apparent is the idea of true long-termism. Prior to 2022, my long-termism felt somewhat insincere. What does this mean? I always believed that what I was doing in my entrepreneurial journey was very ambitious, and I even thought it was significant in human history alongside technologies like Crypto and AI. Perhaps because the first few years of entrepreneurship went so smoothly, although I wanted to strive for long-term, ambitious goals, I was rather impatient during execution.

I remember in 2021, I was very captivated by a book called Blitzscaling. In the prologue, there’s a metaphor suggesting that entrepreneurship is like flying a plane while changing the engines in mid-air after taking off. The goal is to develop rapidly, grow at lightning speed, and expand explosively, disregarding costs to achieve network effects and ultimately become the leader in the industry. Many of our executives received this book as a gift from me.

I bought both the Chinese and English versions for everyone, as I was very taken with it then. It was only after the setbacks of 2022 that I realized while there are indeed correct elements in it, much is also inapplicable. This method itself pursues transformation and innovation but is constrained by the characteristics of the Internet age. One feature of the Internet era's development is that as scale expands, marginal costs decrease, and network effects intensify.

The Crypto industry possesses strong financial attributes, with many opportunities emerging from heavily regulated financial sectors that slowly necessitate shifts in regulations due to these technologies. This does not align with the characteristics of rapid expansion seen in the Internet. In any case, I am re-evaluating this logic now, as it may not be suitable for our entrepreneurial journey.

Regarding long-termism, I now believe that the things we strive for remain immensely ambitious and might require a lifetime to accomplish. For this reason, I don’t overly care whether it gets achieved tomorrow, the day after, or in a year, five years, or even ten years.

Recently, Jensen Huang had an interview where one of his comments resonated with me greatly. He said, "I can spend 50 years changing the world without needing to change the world tomorrow."

This statement resonates with me now, perhaps due to my age. Looking back, a significant error I made in 2022 was actually a strategic misalignment; I incorrectly benchmarked against the wrong reference. I've rarely had a competitive mindset growing up, but in 2022, I felt I did. At that time, my mindset shifted, and paired with the ideology of Blitzscaling, I failed to genuinely uphold the necessary long-termism.

I engaged in rapid expansion and even blind growth. We demanded that products continually deliver new features every two weeks and major iterations monthly. For any business line—from Metaverse, NFT to Payment Card—whatever came to mind, I wanted to do, either by creating it myself or through acquisitions. Having gone through that impatient phase, I realize now that I, along with the entire industry, lost our way during that period. After that experience, I believe now that as long as I can work towards achieving the lofty ideals we should strive for, I don't need to rush; this is the most significant growth for me.

I am a true believer in Crypto, but what I believe may differ from many friends in the industry. I think that as the industry progresses, most people focused on Crypto applications can be categorized into three groups. The first group has fallen into extreme financial nihilism, viewing it simply as a massive gambling platform purely for profit. The second group believes that it will recede into the background, becoming the foundation for traditional financial institutions; I also believe this, but I think Crypto and blockchain applications are not limited to that. The third group might be newcomers or overly technical idealists who believe in pure decentralization, assuming that what is built on top is not just decentralized finance (DeFi), but also decentralized social, and overall decentralized governance, which I believe is somewhat overly idealistic.

My thoughts may differ from these three groups. In terms of the financial attributes of Crypto applications, I agree on these two points: First, it achieves the extreme of permissionless; second, traditional finance needs to collaborate because it brings many benefits of permissionlessness and generates many business opportunities during the regulatory migration process. However, when it comes to non-financial Crypto applications, I believe they hold immense value due to their trustlessness, permissionlessness, and decentralization, which are of significant importance for the future Agent Economy, especially when the mainstream participants in that economy become AI Agents rather than individual humans. However, before that stage arrives, I believe the trade-offs for most human users are not worth it.

Consider it this way: as an individual user, if you have to sacrifice convenience, experience, and costs to use a more expensive, more troublesome, and poorer service that offers the advantages of being decentralized, trustless, and permissionless, most people would be unwilling. Most care about whether things are convenient and affordable—ideally free. However, if today you are an AI Agent, while making a decision, you would definitely consider costs, but that wouldn’t be your sole consideration.

Experience doesn’t matter to you because you are an AI Agent, so convenience is not an issue. But permissionlessness is crucial; if all access requires waiting for human approval, it will be troublesome for an AI Agent. Trustlessness is also vital because if the entire process relies on third-party trust, the outcomes are uncontrollable for an AI Agent. Similarly, decentralization is important; your processes shouldn't be interrupted or altered due to uncontrollable human reasons. Therefore, you will quickly discover that the three advantages of blockchain applications are highly significant for AI Agent users, while their three disadvantages are acceptable to AI Agents. I believe when the entire economy becomes dominated by AI Agents, which I anticipate will be within three to five years, other blockchain applications, beyond financial ones, will generate significant value instantly.

Envisioning the AI Agent Economy

Mercy: What would an AI Agent-led economy look like?

Michael Wu: In fact, we are already beginning to see it. When we call a taxi, it is essentially a ride-hailing app matching us; when we watch short videos, it’s really an algorithm making recommendations behind the scenes. The difference now is that these programs themselves lack commercial decision-making; they are executed behind a company with a predetermined business purpose. That’s why we say for AI Agents, you give any business model a wallet address, letting it value its services with its tokens, and establish payments with humans or other agents. In that case, all decisions become its own.

What would such an economy look like in the future? A company is essentially a network of Agents determining the optimal decision through economic resources priced in Crypto.

Then they can dispatch services; some services are online, just like the applications we use now; others are offline and physical, which will utilize robots. Don’t think that only humanoid robots qualify; even a robot collecting plates in a restaurant, or a door handle with a small chip running a localized model, is an AI Agent. With each service, it first collects data, second dispatches the service, and third charges fees. I believe that by that time, all applications will be agent-driven, and the underlying foundation should be Crypto. In fact, this is already happening; for example, package deliveries are now being done with drones, and Tesla’s FSD has greatly improved its autonomous driving capabilities.

If a large number of cars become autonomous, the person owning the car no longer possesses just a means of transportation, but rather an Asset (asset). What they truly own is not a driver but an asset that generates economic effects. This economic effect might be for personal use or for others. In such an economy, payments will definitely be primarily algorithmic, followed by many high-frequency or even Streaming Payments. Humans generally dislike Streaming Payments, as seen with current many streaming services.

Often, programs are left unfinished; you stop watching after just a few minutes. Over a decade ago, suggestions came up for charging by viewing time, like charging for a minute watched. However, people are averse to this due to perceived pressure, feeling that watching an extra minute translates to spending more. Instead, individuals would prefer to pay more for a bundled subscription, as this aligns with their desires for consumption. But when you actually own an asset, and your experience is decoupled from its usage, presumably the goal would be maximum utility. Traditional payment networks cannot meet the various forms of scattered payments, so at that point, Crypto will definitely come into play.

How Market Makers Provide Liquidity

Mercy: Since you mentioned earlier that you were originally an FX trader, your background is likely connected to Amber’s business today. Can you explain what a market maker is and what they specifically do?

Michael Wu: Everyone knows that Amber is perhaps one of the largest and most successful market makers in the entire Crypto industry. Of course, this is just one of our business lines. The essence of market-making services is Liquidity provision, meaning providing liquidity services. What does it mean to provide liquidity? When someone wants to buy an asset, there needs to be a selling price; and when someone wishes to sell, there has to be a buying price. The optimal situation is that there's always a price available, and you can sell at that price whenever you want. Those who provide this buying and selling price are Market Makers.

There are various market makers; even as an individual, if you are willing to place a buy or sell order, you become a market maker. However, institutional market makers like us use technology to make these services highly frequent, providing significant trading depth to enable assets to trade conveniently and at a lower cost when someone wants to buy or sell. The specific implementation is straightforward, using our algorithms and AI involvement to provide order books that ensure buy-sell spreads are as narrow as possible, thereby minimizing trading costs and maximizing depth.

There are two sayings in our industry that are crucial for the market-making business. The first is Liquidity begets liquidity, meaning liquidity brings more liquidity. The second is Liquidity is most expensive when you need it the most, meaning liquidity becomes the most costly when it is most needed. When an asset's price fluctuates dramatically, determining where its price should be is most challenging. At that moment, as a market maker, if others sell to you, you might be unable to offload; or if others buy from you, you might struggle to repurchase. In extreme market conditions, liquidity essentially vanishes.

We can proudly say that after experiencing so many extreme market situations, our reputation has continuously improved. Because people have noticed that, compared to others, we are still willing to do our utmost to provide liquidity and uphold our commitments as much as possible.

Why Compliance is Crucial

Mercy: I know Amber also places a strong emphasis on compliance. Aside from regulatory requirements, why do you believe compliance is so important?

Michael Wu: I think this is a great question. I’m not a compliance expert, so I can only share some personal views, which don’t represent Amber, but just my own perspective. Why is compliance important? The first point can be for oneself, to prevent compliance risks, which we refer to as long-tail risks. You can choose not to protect against them because most of the time they are not exposed, but once they do emerge, they could potentially ruin a company or cause significant impacts. In our industry, compliance is often challenging, fraught with uncertainties; people believe it won't happen or are willing to gamble that it won’t, but once it does occur, the consequences—whether heavy fines or catastrophic repercussions—are incredibly severe. We have always been a company that values risk management over the years; failure to invest in security could potentially lead to catastrophic theft.

Compliance is similar. We indeed adhere to long-termism, aiming to build something large; this implies that investment in compliance is essential. Looking a bit further, the activities of the entire economy, alongside AI agents and the migration of the AI economy to the chain, represent vast economic interests and impacts on the world and society. Therefore, governments and regulators will inevitably get involved. The sooner we embark on this path, the more likely it is that when the paradigm shift happens, we will have a competitive advantage and moat that allows us to conduct business when others cannot.

Secondly, I believe compliance is a protective mechanism for the entire industry. If one or two companies fail to comply, they may be at risk themselves, but if a large number of such incidents occur, it ultimately undermines regulators' confidence in the entire industry, potentially causing enormous negative impacts. As relatively leading companies, including ourselves, we all share the understanding that we must not wreck the industry. We should at least do our part well. Moreover, we experienced the challenges of 2023, which were extremely painful.

I recall that there was a period when everyone was reluctant to engage with Crypto; it became somewhat of a taboo. Our clients are family offices and institutions, and we had good relationships with them, with their accounts still with us, but in 2023, when we wanted to show them the opportunities in Crypto, they weren't interested. It was only after favorable developments such as the approval of the spot ETF that confidence in the industry began to recover, and the regulatory cycle shifted from a counter-cyclical phase to a pro-cyclical one, allowing the industry to have a chance to move forward again.

Insights into Customer Demographics and the Future of Compliance

Mercy: I just noticed there are many similarities, like the focus on long-termism and pursuing compliance routes. OKX is also going the compliance route. Do you think the client structure has changed at all before and after deciding to pursue compliance? Have their concerns shifted at all?

Michael Wu: Back then, compliance was not really a concept, and achieving compliance was quite difficult due to the novelty of the industry. However, our general attitude remained relatively consistent as our client base has always been concerned about these matters. Many of our institutional or family office clients are unwilling to risk their professional integrity by dealing with a non-compliant organization because if something goes wrong, they become accountable for it.

This has brought some challenges. For instance, we are licensed in Singapore and may soon need to be licensed in other regions as well, which sometimes requires client migration due to regulatory demands. Clients often get frustrated, complaining that they have already onboarded, so why must they re-onboard in another region and submit various materials?

We ventured into B2C business in 2022, but ultimately it didn’t succeed greatly for various market reasons, leading us to shut down B2C operations by the end of 2022. However, I believe in the future, Agents could provide relatively customized services for every client, big or small, especially those on-chain. This includes offering precise solutions for access issues. Through Agents, we can identify clients' fund sources, volumes, purposes, and methods of participation, thereby determining the level of KYC or KYB required for access. We incubated a project this year that aligns with this direction, and I think it's quite promising.

Dealing with Anxiety and Information Overload

Mercy: Firstly, as a CEO, you manage a high density of information daily. How do you filter information and combat anxiety when facing significant uncertainties or pressures?

Michael Wu: I'm someone who responds to messages quickly; I generally hope to process and reply to a lot of information within 5 to 10 seconds. However, if it’s crucial information, I add it to my To-do list or let it settle. This is a habit I developed as a trader; I often have workshops that are personal time for me to tackle specific topics. As for managing anxiety, it’s generally alright; I usually don’t feel much anxiety.

Mercy: So you’re a complete optimist, right?

Michael Wu: Yes, I am pretty optimistic. Sometimes the pressure is high, but as long as I can sleep, it’s okay.

Investment Logic and New Year Outlook

Mercy: I know Amber is now involved in investments and has incubators. When looking at projects, what is most important to you?

Michael Wu: Obviously, the team is very important, and I generally like to have conversations with them. My interests are always in the intersections of Crypto and AI, or they might not directly work with Crypto and AI, but within my vision, if a project has a clear positioning in the Crypto and AI combined world, that would be very attractive to me.

Mercy: Is there a particular aspect that, if a project has it, you absolutely would not invest in it?

Michael Wu: The integrity of the founder. In our industry, defining a rug pull is challenging; if I find out through engagement that there are fundamental issues or if many aspects suggest their values are questionable, I would generally keep my distance.

Mercy: Can you mention something that best represents you and one that you value most?

Michael Wu: It might be a black T-shirt.

Mercy: What are your personal goals for the New Year?

Michael Wu: I definitely want to remain true to my original aspirations and continue leading the company to achieve my grand ambitions.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。