On October 9, at the All-In 2025 Summit, Alibaba Group co-founder and Chairman Joe Tsai shared his latest insights on US-China tech competition, Alibaba's strategic transformation, and AGI among other hot topics. Tsai stated bluntly: "AI is not a winner-takes-all field. True victory lies not in developing the most powerful AI models, but in achieving faster technological implementation and applications."

Tsai engaged in a debate with David Sacks, dubbed the "White House AI Czar" and head of White House AI and cryptocurrency affairs. Sacks emphasized that the US must maintain leadership in AI and chip technologies. However, Tsai argued: "The US should invest more resources in technology applications and popularization, rather than focusing solely on technology R&D."

Tsai added that China actively embraces open-source principles, and many companies have launched models with moderate parameter scales, a development model more conducive to rapid technology adoption. While he did not assert that China has taken the lead in the model technology race, he stated that "China has indeed made significant progress in practical applications." He concluded that AI is becoming a ubiquitous fundamental element like air, and no country would claim exclusive ownership of air.

Regarding Alibaba, Tsai mentioned joining due to Jack Ma's leadership, deeply attracted by Ma's clear vision and ability to identify talent. Throughout Alibaba's 26-year development journey, Tsai witnessed Chinese companies transition from a "free expansion period" to a "new normal," believing that today's regulatory environment is more predictable and stable.

Alibaba has undergone internal transformation in recent years. Tsai revealed that after returning as board chairman, the first thing he did was tell everyone that promoting Alibaba's involvement in six different businesses to the outside world was "too confusing." Tsai focused Alibaba's strategy on two core areas: e-commerce and cloud computing, deeply integrating AI into all business segments, with this focus being key to managing the company well. Currently, approximately 30% of Alibaba's code is generated by AI.

Regarding AI's future, Tsai specifically mentioned Google DeepMind founder and CEO Demis Hassabis's prediction that AGI might be achieved within 5 to 10 years, but Tsai personally believes AGI still needs about 20 years to truly arrive.

Beyond Alibaba, Tsai has another identity: he owns NBA team Brooklyn Nets and WNBA team New York Liberty. In the interview, he also shared business insights from operating teams and perspectives on the sports industry.

Recently, the 2025 NBA China Games will be held in Macau, marking the event's restart after a 6-year hiatus. The participating teams are the Phoenix Suns and Brooklyn Nets, with the latter being Tsai's NBA team, while Alibaba Cloud also sponsors the event. On October 9, NBA China and Alibaba Group announced a multi-year partnership between NBA China and Alibaba Cloud.

The All-In 2025 Summit was held from September 8-10 in the US, initiated by the well-known Silicon Valley podcast All-In. This year's summit also featured notable tech industry leaders including Elon Musk and Demis Hassabis.

The following is a transcript of Tsai's interview at the All-In Summit (with team-related Q&A moved to the end):

**01. Alibaba Founded 26 Years Ago, Regulatory Environment Has Entered New Normal**

Host: Let's talk about Alibaba. This has been an incredible journey, with Jack Ma's story documented in books, but can you take us back to your mindset then? How did you get involved, what were you thinking, and how much risk did you take?

Tsai: I was fascinated by Jack Ma's personality. People see Ma's public side, which is very charismatic. But what I saw was his ability to instill belief in people. When I walked into his apartment, there were about 12 to 15 recent college graduates, and Ma was like a teacher 10 years older than them. He was good at communication and could paint a very clear vision. That's what I truly identified with and joined for. His leadership ability was amazing.

Ma himself came from a teaching background, was educated as such, and taught English at Zhejiang University. Teachers are natural leaders to some extent. Teachers must first be good communicators, and secondly, they must be able to identify talent. I think many teachers enjoy making judgments like "this child will be very successful in the future," writing recommendation letters for them, doing various things. And teachers are humble enough; they are willing, even happy to see their students more successful than themselves. In the process of starting a company, you need to be able to accommodate all kinds of people who are smarter than you. This is very important.

Host: There was a phase, I think around the early 2010s, when China felt like a highly free capitalist environment, with a Cambrian explosion of excellent entrepreneurs emerging, including you, Jack Ma, Pony Ma, and others. Later, things changed somewhat. Can you outline this process of change for us? I believe these changes profoundly affected foreign direct investment entering China's ecosystem. Please tell us about this journey you personally experienced, participated in, and observed.

Tsai: Alibaba has been established for 26 years. I think the first 15 years were completely free market growth, where we organically built many businesses. Then we entered a phase of extreme competition. Everyone wanted to do e-commerce because if you have online traffic, e-commerce is the best way to monetize that traffic. In today's market, we have five or six very strong competitors, including ByteDance, TikTok's parent company. They're also doing e-commerce. They're not known for e-commerce, but they're one of our fiercest competitors.

We went through a period of extreme competition, then the government felt the industry was somewhat off track. Competition was too intense, and some platforms showed monopolistic behavior. So more regulatory measures were introduced, some of which were actually very good, such as privacy protection and anti-monopoly measures. Now we've entered a new normal where the regulatory environment is more predictable. We know where the red lines are, what should and shouldn't be done. Actually, due to this predictability, it has created a better business environment.

**02. China is a Peaceful Country, Emphasis on Education Worth Learning from by America**

Host: On the geopolitical level, the US seems to constantly indoctrinate Americans to view China as an existential threat to America, that we are competitors, that we cannot cooperate and jointly lead the world toward prosperity. What's your view?

Tsai: I disagree with that worldview.

Host: I don't believe it either. I don't know why our only tone of dialogue has to be that we are mortal enemies.

Tsai: I think I can understand why America has such thoughts. China has risen rapidly over the past 20-25 years, manufacturing has become strong, becoming an export powerhouse. Due to economic development, China has also become a tech power. America's fear is that economic strength and technological strength will translate into military strength, making it a national security issue. I understand this.

But I think China and the US are the world's two largest economies. I think you need to step back and think about the problem. On one hand, you should compete with China. Hyperscale companies compete globally with Chinese internet companies. But on the other hand, there are so many conflict zones in the world. Didn't President Trump say he himself prevented seven wars? China didn't start any of them. How many years has China not started a war? I feel it's been 30-40 years. I think China participated in the Korean War, maybe the Vietnam War, but that's ancient history.

If you observe the Chinese people and China's behavior, it's a very peaceful country. China cares about its own economic development and the welfare of its citizens. I think there's a lot of friction in the US-China competition process, with China feeling that America is trying to contain China and prevent China's economic rise.

Host: Speaking of citizen welfare, you mentioned ByteDance. I think there are reports that ByteDance's revenue just exceeded Meta's revenue. ByteDance's core product TikTok is fundamentally different in America and China. Due to regulation and concern for citizen welfare, China's Douyin has strict regulations on content, screening, and algorithms. In America, TikTok is full of cat videos and all kinds of meaningless content, although I now notice TikTok has added a STEM section. China has that mindset and willingness to think about its citizens. I'm just curious, how should we Americans embrace more of that principle? Because this is often overlooked in the US-China confrontational atmosphere. What should we do more of that China is doing?

Tsai: Education. China has a highly educated population. Talk to any Chinese parent, and they want their children to take the gaokao and enter the best universities. If they want to send their children abroad to study, their targets are Ivy League schools or Stanford. Chinese people place great emphasis on education. I think America has some structural problems, like teacher union obstacles. But China doesn't have teacher unions.

**03. AI Field is Not Winner-Takes-All, US-China Still Has Broad Cooperation Space**

Host: David, I'm curious how you view this issue. Are we destined to forever be in conflict with China? Or under our remarkable, dynamic President Trump's leadership, is there another path forward?

David Sacks: America and China are in high-tech competition, also competing in security and economic fields. The reason is that China has become wealthy and powerful, and America doesn't tolerate peer competitors. America wants to be the strongest country. Power balance is somewhat a zero-sum game. Economics isn't, but power is. America's history is that we want to be first, we don't like having pure competitors. That's the bottom line. This is for good reason. Countries are essentially in anarchy with each other, meaning there's no higher authority above nations. If you encounter trouble in the global system, you can't dial 911. Countries place survival above all considerations, and the way to survive in the international system is to become powerful. Countries measure your power by the gap between you and the next most powerful country. 20 years ago, China wasn't seen as a threat because it wasn't wealthy and powerful, but now it is.

I think this has led to a more hawkish atmosphere in Washington. I agree with parts of their view because I think it's very important for America to win the AI race. We don't want China to dominate AI or the chip field. As Americans, we want America to be the strongest country. If I were Chinese, I would want China to be the strongest country. I have no personal grudge or malice about this. But I want America to be the strongest country, and in my little field, this means winning the AI race.

Tsai: May I respond?

Host: Of course. Please go ahead.

Tsai: I understand your feelings as a patriot wanting America to win. But regarding the AI field, I don't think there's such a thing as winning the race. It's more like a long marathon. Observing the development trends of model companies, you'll find new models leading every week, but the next week they're overtaken by other models. On the other hand, the entire AI industry hasn't yet formed a clear business model. The e-commerce business we're in has network effects market models, but I'm skeptical whether the model development field also has this winner-takes-all characteristic. I think AI is not a winner-takes-all field.

In my view, true victory lies not in developing the most powerful AI models, but in achieving faster technological implementation and applications. America should invest more resources in technology applications and popularization, rather than focusing solely on technology R&D. Currently, each tech giant invests about $80 billion annually in R&D, which is a huge investment.

Looking at China's development path: First, China actively embraces open-source principles; second, many companies including ours have launched models with moderate parameter scales - such as 1.7 billion and 8 billion parameter models, which are more suitable for mobile devices and laptops. This development model is more conducive to rapid technology adoption.

I should clarify that I'm not asserting China has taken the lead in the model technology race, but China has indeed made significant progress in practical applications. A survey last year showed only 8% of Chinese companies applied AI technology in their business; now this proportion is approaching 50%. The speed of this application adoption is remarkable.

AI is becoming a ubiquitous fundamental element like air, and no country would claim exclusive ownership of air. In critical fields like healthcare and biology, AI holds enormous potential, providing broad space for cooperation between the two countries.

**04. After Returning to Alibaba, the Key is Focus**

Host: Whether self-developing or adopting existing models, the most critical question is: what is AI's actual impact on business operations? How will this change hiring scale? At what levels will it have impact? What changes will it bring to efficiency and work nature? What are the specific AI application scenarios within Alibaba?

Tsai: AI has indeed significantly improved our operational efficiency, correspondingly reducing manpower needs. We're trying to use AI to write quarterly financial reports, even considering using AI in analyst meetings, observing whether people can tell the difference, then we'll just lay off the entire finance department (laughs). But AI's greatest value lies in empowering consumer-side applications. We deeply integrate AI into e-commerce, maps, food delivery and other businesses, expanding user scale by improving user experience, with this impact ultimately reflected in revenue growth.

Host: As more and more work is done by AI, won't demand for labor naturally decline?

Tsai: We haven't announced any layoffs because of AI yet, but I keep asking our engineering department heads how much code is now written by AI. I get different answers depending on which department you ask, but I think it's probably reached 30% now.

Host: Already that high?

Tsai: Yes, I did a weighted average across different departments, roughly that figure.

Host: Alibaba is a huge enterprise with business spanning many countries globally. How does Alibaba build corporate culture, and how do you manage massive numbers of employees, diverse needs, and various issues? How do you handle this?

Tsai: Focus. Before taking over as chairman, I was actually gradually stepping back from company operations, focusing on the Brooklyn Nets. About a year and a half ago, I returned as company chairman. The first thing I did was tell everyone we can't keep saying our company is involved in six different businesses - it's too confusing. We only do two businesses: e-commerce and cloud computing, both containing AI elements, these two are our core businesses. This focus allows our team to concentrate and execute company decisions.

Host: Finally, can you briefly predict the team's performance this new season?

Tsai: Are you referring to the entire league, or specifically the Nets?

Host: Let's talk about the Nets.

Tsai: We're in a rebuilding cycle. This summer we fully utilized five first-round draft picks while retaining important draft assets for 2026. This layout already indicates our strategic direction - we'll focus on developing young players, giving them ample growth space. While this means we need to go through a growth process, this is our clear development path at this stage.

**05. AGI Still Needs 20 Years to Arrive, China Doesn't Fear Superintelligence Getting Out of Control**

Host: I want to follow up with a question. We just talked in the previous panel discussion that China seems to have a somewhat different view of autonomous driving technology. China is both testing autonomous driving technology but also wants to ensure related technology won't cause widespread unemployment. What exactly is China's view on autonomous driving? Tens of millions of workers make a living driving - what are their views on this? I understand some drivers in Wuhan have expressed dissatisfaction. How do the Chinese government and Chinese people view this incredible transformation of autonomous driving and its impact on employment?

Tsai: I think the government fully embraces it. The Chinese government introduced the "AI+" policy a few weeks ago - by 2030, five years from now, they want to see 90% adoption of AI agents, devices, etc. in society. This is an official public government statement - full commitment. There really isn't much discussion about AI replacing human jobs. But overall, Chinese people have employment anxiety. China has about 10 million university graduates annually. China's youth unemployment rate, that is, the unemployment rate for the 16-24 age population, including graduate students, is 18%. Actually quite high, and Chinese people have a lot of anxiety about this.

Part of the reason is economic. Although China's overall economic situation is quite good in terms of infrastructure, energy access, etc., there's currently a kind of low mood because China has experienced a real estate market decline over the past four or five years, with average housing prices falling about 30%. This is a very negative wealth effect, and this impact continues.

David Sacks: Does China fear AGI? I'm talking about fear of superintelligence getting out of control. Is this a widespread fear, or just an American concern?

Tsai: Government insiders don't talk much about this because they believe they can control AGI better, perhaps better than America. But there are indeed some lingering concerns. I just heard Google DeepMind founder and CEO Demis Hassabis speak, his prediction for AGI is 5 to 10 years. I think AGI might take another 20 years. The key is the word "general." AI must be able to generalize and apply principles to scenarios it has never seen before.

**06. Teams are Both Business and Social Institutions, NBA is Essentially a Product**

Host: You own the Brooklyn Nets (NBA team) and New York Liberty (WNBA team). Operating teams must not be easy either?

Tsai: Yes, not easy at all. Any professional team is both a business and a social institution - we need to keep fans happy. There are some church and state elements in this, no different from running a news organization.

Host: Can we start by talking about the WNBA? Recently Caitlin Clark's season ended, causing a lot of controversy and attention. What did the WNBA do right? What did it do wrong? Did the league fulfill its responsibility to protect this star player?

Tsai: Caitlin Clark undoubtedly had a huge impact on the WNBA. All metrics are rising. In her first season entering the league from college, we saw viewership, ticket sales, sponsorship revenue - almost all metrics quadrupled. Her economic impact on the league is extraordinary.

Host: Indeed undeniable.

Tsai: But because of her characteristics, she's a different kind of player - a point guard, relatively small, and there were questions about whether she could adapt to the league's physical confrontation, but she proved she could. Of course, there's always competition, like her rivalry with Angel Reese that started in college. Some people portray it as racial confrontation, but I think we should move beyond that and focus on the players' skills and talents themselves.

Now there are more and more excellent college players in the league, like Sonia Citron, who's now a hot candidate for Rookie of the Year and broke the league's three-point shooting percentage record. If you ask what the WNBA did right, I think it was really the confluence of timing, geography, and people. But what really changed was that before the season Caitlin Clark appeared, the fan base watching ESPN basically didn't watch women's basketball. Maybe some people followed college games, but definitely not the WNBA. But now, mainstream sports fans like you all are watching the WNBA.

Host: I have an observation about this. You and I are both old-school basketball fans who watched games in the '80s and '90s, watched Patrick Ewing and Charles Barkley. Currently, the WNBA has more intense physical confrontation, they're tougher. Every time LeBron jumps and gets fouled, he flops around. So maybe you can comment on LeBron, or the NBA as a "product."

Tsai: The NBA product is great. I think comparing the physical confrontation intensity between men's and women's leagues is unfair. Both are very physical. If you sit courtside, what you see is physical confrontation and athleticism, right? But games have other dimensions. I now serve on the NBA's "competition committee." Each season, the committee makes rule adjustments to make the NBA product better. When I got the call, I thought, you want me to sit on a committee with Jason Kidd, Chris Paul, Coach K? Can my understanding of basketball really be deeper than theirs? I told NBA Commissioner Adam Silver this committee should be called the "product committee" because rule adjustments essentially determine the game "product" audiences see. For example, changing three-point line position isn't just a game rule issue, it's more a "product design" issue.

Host: When you see the Boston Celtics make 43 three-pointers in a game, do you think that's a good product?

Tsai: I think it's a great product because there are 18 different ways to make 43 three-pointers in a game.