Dialogue with Weijian Shan: Good CEOs make huge fortunes, bad CEOs get out.

Text by Hai Ruojing

Mergers and acquisitions are often thrilling, yet most of their stories are kept confidential. Once the non-disclosure period expires and the details can be shared publicly, the historical context has usually faded with time, with many details lost over the years. Shan Weijian wrote down several transactions he personally experienced, making the conflicts, compromises and enterprising efforts in M&A no longer vague, turning them into concrete, well-documented stories with real people and clear interactions.

At the end of 2004, when the acquisition of Shenzhen Development Bank led by Shan Weijian was at the final critical stage, the team discovered a suspicious loan as high as 1.5 billion yuan: the fund turned into bad debt almost immediately after it was issued. At that time, the net capital recorded on the books of Shenzhen Development Bank was only over 4 billion yuan, and this single loan accounted for 1/3 of the bank's capital.

Who profited from this "huge and bad" loan? The truth was never fully clarified. Shan Weijian wrote in the book that everyone they tracked down complained that they were the victim.

This was just one twist in the Shenzhen Development Bank acquisition case. Eventually, this transaction delivered an impressive performance: Newbridge Capital's initial investment of 150 million US dollars generated a net cashout of 2.25 billion US dollars six years after exit, marking a 15x return.

The negotiation lasted for 30 months. Shan Weijian once rejected the seller's offer of 1.5 times the book net asset value, but later accepted a higher price of 1.65 times; the first foreign president selected after the acquisition knew how to run a good bank, but had no experience turning around a bad bank, and the team had to replace him even after the China Banking Regulatory Commission was alerted.

Bad debts like this "huge and bad" loan were only one of the many non-performing loans Newbridge discovered. Eventually, they found suitable managers, reformed the compensation system, and made bad debt recovery a highly sought-after internal business. Shan Weijian told us that his years of experience made him deeply realize that the role of people is incredibly critical. A good CEO can turn a bad enterprise around, while a bad CEO can ruin a good enterprise.

The understanding of people runs through the entire process of M&A negotiations and post-acquisition restructuring. What the counterparty cares about, what they worry about, what terms can be fought for, and when it is appropriate to compromise, directly determine the success or failure of the transaction. Facing price disagreements, Shan Weijian once did not want to make concessions, but his boss David Bonderman made a prompt decision, pointing out that the 25 million US dollar price gap was insignificant, and the key was to seize the opportunity to close the deal.

You do not have to push negotiations to the absolute limit. Leave some leeway for others, even if they are not willing to do the same for you.

The Shenzhen Development Bank acquisition was the iconic starting point for Shan Weijian's heavy investment in China as a private equity investor. After exiting the transaction, he raised 2.5 billion US dollars and co-founded PAG, creating multiple controlling stake transactions, typical examples including Wanda Commercial Management, Yingde Gas, and Ocean Music which was later merged into Tencent Music.

Nowadays, many types of financial transactions are moving towards quantification, and AI is increasingly adept at analysis and calculation. However, in a complex M&A transaction, there are too many competing forces, commitments may change, information is always incomplete, and different parties often have different understandings of the same plan. Finding opportunities in deadlocks requires human wisdom and imagination especially. Back in 2018, Kai-Fu Lee listed M&A experts as one of the professions that AI is the hardest to replace, which still holds true today.

In mid-2026, when we met Shan Weijian, he was carrying a heavy black shoulder bag alone, with worn edges and corners. He talked to us about the stories in his new book, and spent some time discussing AI.

This investor who grew up in the Gobi Desert of Inner Mongolia completely disagrees with the "AI doomsday theory". He wrote a book review refuting Yuval Noah Harari's statement in *Homo Deus* that AI may rule humans. In his view, AI will surpass human intelligence in many aspects, but it has no "survival and reproduction needs", so it will not generate internal motivation or autonomous consciousness.

"When Einstein was a teenager, he imagined traveling synchronously with light. In his thought experiment, light did not move at the speed of light, and was relatively stationary with him, just like people sitting on a plane today cannot feel that the plane is moving at high speed. I think this kind of imagination is unique to humans."

His confidence in human intelligence did not make him underestimate the impact of AI on business and social organizations. Earlier this year, because it was difficult to judge AI's impact on the target company, they chose not to invest in an IT service company.

As for the upcoming unemployment wave and extreme wealth gap brought by AI, Shan Weijian is more optimistic about China's situation. He believes that China has the ability to cope with technological shocks through distribution systems and social security mechanisms.

During the conversation, Shan Weijian always brought the topic back to "what humans can do" to respond to changes. The following is the edited conversation between Waves and Shan Weijian:

 

Oppose AI doomsday theory, but stay alert to extreme wealth gap

Waves: Earlier this year, you wrote a commentary refuting the "AI doomsday theory" mentioned in Harari's book *Homo Deus*. You believe that AI does not have the irreplaceable imagination of human beings, and it is impossible for it to rule humans in the future?

Shan Weijian: I do not agree with his core view, the so-called AI doomsday. He believes that artificial intelligence can replace human intelligence, and even rule humans. I think this is impossible.

The most fundamental foundation of human intelligence is the need for survival and reproduction, which is the root of intelligence development. All human activities are determined by this fundamental need. Therefore, humans can design behavioral goals.

For example, the fact that you and I are sitting together today is a goal we set. Everything we do is an autonomous decision, which requires people to be aware of their own existence, have self-awareness, and know the goals they want to achieve.

AI can develop to be stronger than human intelligence in many aspects, but it will not have autonomous will. Without autonomous will, it is impossible to set goals independently. The goals it sets are still the goals that humans require it to achieve. So it cannot control human beings.

In the movie *The Matrix*, AI enslaves humans and designs the purpose of extracting human energy. I think this is absurd. How much energy can humans generate? (Laughs) If AI's capabilities have reached that level, it can completely generate energy through nuclear reactors and other methods. So what purpose would the so-called AI doomsday of controlling humans serve?

Waves: This is more like a reflection on philosophy of technology. You believe AI has no motivation.

Shan Weijian: It has no motivation, no purpose. It does not need to reproduce, nor does it need to survive, so I think the so-called claim that artificial intelligence will eventually replace human intelligence and even control humans is alarmist and impossible to happen.

I did not form this opinion when I wrote this book review. More than a decade ago, physicist Stephen Hawking expressed the concern that "AI will eventually control humans", and I thought about this issue at that time. A few years ago, I published an article stating that AI will surpass humans in many aspects, but it cannot control humans because it has no imagination.

AI can reach the goals set by humans faster, it can make extensions, but it cannot imagine something that does not exist in the training dataset out of thin air. When Einstein was a teenager, he imagined traveling synchronously with light. In his thought experiment, light did not move at the speed of light, and was relatively stationary with him, just like people sitting on a plane today cannot feel that the plane is moving at high speed. I think this kind of imagination is unique to humans.

Waves: At present, AI has not controlled humans, but it has already brought huge impacts to the business world, such as layoffs, and many people are anxious that their jobs will be replaced by AI.

Shan Weijian: Which industrial revolution did not bring huge impacts to the human world? The first industrial revolution, when the steam engine replaced human labor, did not cause human extinction, but brought further prosperity. The electrification revolution and the computer revolution are the same.

But artificial intelligence will create a big problem: wealth concentration. This has been vividly demonstrated in the United States, where it has created trillionaires and is causing many people to lose their jobs (such as Google's large-scale layoffs). The uneven wealth distribution is further intensifying the polarization between the rich and the poor.

Waves: How should we respond to unemployment and extreme wealth gap?

Shan Weijian: I think China has advantages in this regard, and its distribution system will not be so extreme. Over the past few decades, one of China's great achievements is poverty alleviation, allowing some people and regions to get rich first, and then drive others to achieve common prosperity. The TV series *A Land So Rich in Beauty* truly reflects one of China's poverty alleviation methods: rich provinces pair up to support poorer provinces. This can only be achieved under China's system. Some people in the United States also want to do the same, but there is no systematically planned distribution arrangement at the institutional level.

Waves: In the AI era, there may not be so many jobs needed? Many companies are recruiting fewer new people now.

Shan Weijian: Every industrial revolution will bring such a phenomenon, some people's mastered skills become outdated, for example, AI replaces humans in writing code.

In the early years of the state-owned enterprise reform, many workers were laid off during the restructuring process. At that time, the government designed a re-employment training system to guarantee the basic salary of workers in the transition period. In the early 1990s, the social security system was established, and later a national security system was set up. These institutional designs gave laid-off workers the opportunity to find new jobs.

I think China's system has certain advantages. It will not abandon the unemployed after AI replaces a large number of human jobs. Moreover, current social conditions are better, and workers have strong self-improvement and retraining capabilities.

Waves: Specifically for PAG's investment decisions, is AI becoming a key variable for you to judge a company?

Shan Weijian: It depends on the type of project. Earlier this year, the PAG team looked at an Indian IT service company. India has many highly educated, English-speaking employees with low labor costs, and European and American enterprises generally outsource software and IT services to India, which used to be a very profitable business.

At that time, the trend of AI replacing programming and IT services was not very clear. At the investment committee meeting, we asked: if AI completely changes the way of programming, what impact will it have on this company? Several experts invited by the team basically believed that the impact would be positive: for example, the work that originally required 100 people can now be done by 10 people.

But we thought further: European and American customers choose India because of low labor costs. If AI is cheaper than Indian labor, customers may no longer need outsourcing. They can just tell AI their goals and the software can be developed.

So, will AI help this company improve human efficiency, or will it make customers no longer need its services? At that time, we could not judge, so we did not invest in the end. Considering the future exit, if potential buyers also cannot judge AI's impact on the company and are uncertain about the company's future value, no one will dare to acquire it.

I now think that many software companies will hardly find buyers at almost any price, because it is still difficult to predict the impact of artificial intelligence on their future.

 

Negotiations do not need to be pushed to the extreme

Waves: In *Money Games: A Story of Twists and Turns*, the first point that impressed me deeply is the price game during Newbridge's acquisition of Shenzhen Development Bank. Newbridge believed that the reasonable price of Shenzhen Development Bank was 0.8 to 1.5 times the book net asset value. The counterparty proposed 1.5 times, which you rejected, but you reviewed afterwards that "it was a wrong move that almost caused the whole deal to fail". Later, the two sides returned to the negotiation table, and you finally accepted the price of 1.65 times. How was the disagreement resolved?

Shan Weijian: At that time, the book net asset value of Shenzhen Development Bank was not real, and there was an asset gap in principle, so we thought that 1.5 times the book net asset value was already the upper limit. The counterparty insisted on 1.65 times, which I found very difficult to accept. But my former boss David Bonderman said that the 25 million US dollar difference was not a big deal and insignificant, the most important thing was to close the transaction.

He taught me that when negotiating, you cannot push it to the absolute limit, you have to leave something for others. Even if others may not leave you anything, you should leave some leeway for others.

Waves: Similar trade-offs often appear in M&A negotiations?

Shan Weijian: When Newbridge sold Korea First Bank, we negotiated with Standard Chartered and HSBC at the same time. The two offers differed by more than 100 million US dollars, and the Korean regulatory authority tended to choose the bidder with the lower offer.

100 million US dollars is a huge number for ordinary people. But in Bonderman's view, it was just a "rounding difference", and we still needed to consider the Korean government's preference.

Until later, the bidder with the higher offer substantially raised its bid, making the gap very meaningful. The Korean government was a 49% shareholder of Korea First Bank, and the price increase also had a substantial impact on them. Eventually, they also accepted the higher offer.

Waves: It sounds like a very simple common sense: you need to think in other people's shoes and consider their positions and interests. There are many parties involved in M&A transactions. What is the key for everyone to build trust?

Shan Weijian: Any trust takes time, and I think the most important thing is to keep your promise. You must act on your words and deliver what you say. But when others do not keep their promises, you also need to accept it. No one is perfect, everyone makes mistakes.

Keep your word, put yourself in others' shoes and consider their interests, and trust will gradually be built naturally.

Waves: If others go back on their words, or do not recognize the concessions you made from their perspective, it will inevitably make people unhappy. How do you deal with disputes encountered in negotiations?

Shan Weijian: Su Dongpo wrote that "do not panic when something unexpected happens, do not get angry when you are wronged for no reason". I have always used this standard to measure myself. I can achieve the first half, but the second half is very difficult.

I wrote in *Money Games* about my negotiation experience with Korean officials, when we discussed the valuation of non-performing assets of Korea First Bank. We made concessions and proposed a plan with a higher price than it should be, but the other party dismissed it and thought the non-performing assets were high-quality assets. That made me unhappy. But a colleague who participated in the negotiation reminded me afterwards that everything we did was work, and we should not let it affect our personal emotions.

For many things, if you stand at a higher and more distant perspective, and think back after stepping away, you will find that it is no big deal. So do not bring emotions into negotiations. It is very difficult to achieve "do not get angry when you are wronged for no reason". I actually cannot do it, but I will try my best.

Waves: After Newbridge took over Shenzhen Development Bank, the restructuring process was very tortuous at the beginning, and there was a storm of replacing the president. You also replaced personnel when you acquired Korea First Bank before. Do the new management need to be more tough in controlling M&A transactions? As a management professor, what do you think of the combination of firmness and flexibility in management?

Shan Weijian: Management professors do not necessarily know how to manage. Especially when I taught management, I did not understand management at all (laughs).

One important reason why private equity acquisition transactions are more likely to succeed than fail is the management of the invested enterprises.

Our feature is that after acquiring the enterprise, we will provide the management team with a very good incentive system. Through stock options, co-investment and other methods, the interests of both parties are closely tied together, the company and shareholders make