Tencent Still Has a Dream

Original Title: "Tencent Still Has Dreams"
Original Author: Observationist Beating

On August 12, 2026, Tencent released its second-quarter financial report.

Capital expenditure for the quarter, $52.78 billion. Three months ago, this number was $31.9 billion. Looking back a year, the total for the whole year was only $79.2 billion.

The company has always been known for its restrained spending. Its pace of acquisitions was once so slow that the market doubted whether it really wanted to join the AI game. Now, it has ramped up its spending to this level within a year.

On the same day's earnings call, Ma Huateng said Tencent is building a "brand new, AI-empowered Tencent." Melding has been renamed HY, and Hy4 is about to be released. The last time this company described itself as "brand new" was in the era of WeChat's birth.

Tencent still has dreams. Its dream is not just AI; it wants to relearn how to be an unstable company.

In 2018, Pan Luan said in "Tencent Has No Dreams" that Tencent is like water. Water benefits all things but does not compete. Where there is a channel, it flows. Water has no personality, so water has no dreams.

Water naturally flows downhill; this is its nature. Swimming against the current is the business of those who swim against the current. In 2026, this 28-year-old company did something against its nature. It admitted that the article eight years ago was right, acknowledging that living like water is no longer sustainable.

Water

On May 5th at 9 p.m., "Tencent Has No Dreams" was published. Thirteen thousand words.

At 2 a.m., Liu Chiping and Tencent's PR Director Zhang Jun responded on Moments. Liu Chiping said Tencent is a larger organization and ecosystem than the outside world imagines, "simplifying Tencent into a product's successes and failures, a strategic deployment, or one person's will is too narrow."

At 2:19 a.m., a screenshot allegedly from Ma Huateng's response started circulating on Moments.

At 2:39 a.m., the real Ma Huateng finally spoke, saying to a friend who cared about him, "Criticism is good."

He then said:

"From writing the first line of code, my ideal has always been how to create the best product, not how much money to make."

During the day, the national media almost entirely quoted that screenshot that was suspected to be a response from Ma Huateng. Even Zhang Yiming spoke up for Tencent on his Moments, saying this was a "Don Quixote-style imagination," stating that Tencent is not only powerful but also continuously evolving in all aspects.

Zhang Jun was on a long-haul flight that day. After landing, he said, "Indeed, we are not as bad as the outside world thinks, but the criticism has also made us realize that we are not as good as we thought."

Of course, at that time, there were also some different voices about that article. Hong Bo said that many of the issues mentioned in the article are indeed real, but for such a large company, is there a single correct answer? "Perhaps the author believes that Zhang Yiming is the only correct answer. He is a bit superstitious about Zhang Yiming."

The article also recounted an earlier story. In early 2011, shortly after the end of the 3Q war, Tencent held an executive meeting to discuss what Tencent's open capabilities were. Ma Huateng asked the 16 executives present to write down on paper what they believed to be Tencent's core capabilities, resulting in a total of 21 answers. In the end, two were finalized.

Capital, Traffic.

The term "Capital" was advocated by Liu Chiping. Openness means releasing traffic and turning it into investment. Opening up traffic, opening up capital, "I no longer do it myself."

These two words steered Tencent for a decade. WeChat managed the entrance of traffic, investments managed the exit of traffic, with games and advertisements in between constantly generating cash. JD's e-commerce entrance joined the WeChat nine-grid, Sogou took over the search, and Meituan took over local life. Traffic exchanged for equity, equity exchanged for allies. In a decade, Tencent's market value increased tenfold, surpassing Facebook. At the time the article was published, it still seemed invincible.

Looking back eight years later, rereading the article would reveal that it almost predicted every stumble Tencent would later take.

On December 23, 2021, ten years later, Tencent distributed 14.7% of its JD.com shares to its shareholders, valued at approximately HK$100 billion. In January 2022, it reduced its holdings in Sea, cashing out $3.2 billion. In November 2022, it distributed another 9.6% of Meituan, approximately HK$159.4 billion.

The capital that was designated as a "core capability" back then was dismantled by Tencent itself. After flowing for decades, the water began to flow back for the first time.

The First AI Dream

In that article, there was a section that few people paid attention to back then. It mentioned Tencent's AI. The AI Lab's Go program "Masterpiece" was defeated by two amateur works. One was a personal hobby of a ByteDance Vice President, and the other was a side project of a few engineers from the WeChat translation team.

Few people realize that Tencent's first AI dream had already begun two years before the publication of that article.

In the spring of 2016, when AlphaGo defeated Lee Sedol, the world took its first serious look at artificial intelligence. Chinese internet companies seemed to hear a starting gun. That year, Tencent Vice President Yao Xing established the AI Lab. The following year in October, Alibaba's DAMO Academy was founded, announcing a three-year investment of $100 billion.

On March 23, 2017, Tencent announced the appointment of Zhang Tong as the director of AI Lab. Zhang Tong is one of the top Chinese scholars in the field of machine learning and was previously the Deputy Dean of Baidu Research Institute. Upon joining, he said, "Chinese talents worldwide have a strong technical advantage in the field of AI, which is China's opportunity to develop AI."

At that time, AI Lab had over 50 scientists, with over 90% holding AI-related doctoral degrees and overseas study backgrounds. Not content with this, in May 2017, Tencent appointed Microsoft Research's Chief Researcher Yu Dong as the deputy director of AI Lab, in charge of the newly established Seattle laboratory. The reason for setting it up in Seattle was that many Microsoft employees were unwilling to leave the area, so Tencent set up the laboratory next to Microsoft.

The opening of the first dream was impressive. In March 2017, Jueyi won the UEC Cup in Japan, and at the end of the year, it climbed to the top of the Longxing War. Go was the calling card of AI in those two years. After AlphaGo, every AI had to be able to play Go, and that was also Tencent AI's shining moment.

Then, the dream began to deflate.

The problem lay in Tencent's own approach. On September 30, 2018, Tencent initiated the 930 reorganization and established CSIG. Three days after the reorganization, Caijing published an article that directly pointed out the issue:

“Three AI teams still operate separately. How can the B2B business break the deadlock of competition?”

AI Lab, Youtu, WeChat AI—each operating independently, with no data exchange or talent sharing.

On January 3, 2019, Zhang Tong stepped down from his corporate management position and returned to academia. He later went to the joint laboratory of Hong Kong University of Science and Technology and Innovation Works. Over 50 scientists dispersed, and Zhang Zhengyou took over as director.

In 2021, Yao Xing also left. Instead of joining any major tech company, he founded his own company, Yuanxiang, focusing on the "full-stack internet." Among the angel investors is Tencent and Hillhouse Capital.

In December 2025, Yu Dong, the head of the Seattle lab that year, left the AI Lab and joined Capital One in the United States. Three months later, on March 21, 2026, Tencent officially disbanded the AI Lab, with some personnel joining the Hybrid Big Model Team.

From Zhang Tong taking office to the lab's dissolution, nine years. A complete lifecycle of a dream.

Tencent's first AI dream did not fail to launch, nor was it ignored. It recruited the most expensive talent in this field and built the lab next to Microsoft. It perished on a simpler problem: the company didn't know how to nurture a dream that couldn't be monetized immediately.

Research remained theoretical, the business lacked support, three lines raced against each other, with no one losing, so no one winning.

Dismantling

On September 30, 2018, five months after the publication of "Tencent Has No Dreams," Tencent launched the 930 transformation. MIG was disbanded, PCG and CSIG were established, and the six major business groups took shape. Pony Ma wrote in an all-staff letter:

“Rooted in the consumer internet, embracing the industrial internet.”

No one could prove that the article drove this transformation, but what mattered was the resonance. The issues raised in the article were slowly acknowledged by Tencent itself. This was Tencent's biggest transformation in six years, and Pony Ma said this made him feel “like walking on thin ice every day.”

After the transformation, the newly established CSIG at Tencent, led by Tang Daosheng, announced that CSIG would no longer use the horse racing mechanism.

Horse racing was Tencent's heritage, and WeChat was the product of that. In 2010, several teams were simultaneously developing mobile instant messaging, and Zhang Xiaolong's team in Guangzhou won. This mechanism produced Tencent's last great product.

On November 11, 2019, Tencent's 21st anniversary, Pony Ma sent an all-staff letter announcing a new mission and vision:

“User-Oriented, Technology for Good.”

The old vision of Tencent was to be the “most respected internet company,” grand and vague, with no clear path to realization. A year and a half later, Tencent changed it.

Then there's Tencent Cloud. The industrial internet which had high expectations during the 930 period didn't go smoothly. In 2021, Tencent Cloud stopped pursuing scale and shifted focus to reducing losses. By 2023, Tang Daosheng clearly stated the need to transition from being an integrator to being integrated, scaling back unprofitable overall integration projects. The gross margin of financial technology and enterprise services slowly climbed to 51% in 2025.

The cost is that the ambition of this transformation, which started as "rebuilding another Tencent," has shrunk to "making profits right."

Aside from reducing losses, these five years have also borne fruit. Tang Daosheng coined a term for the Industrial Internet, called C2B. The Industrial Internet is not only ToB and ToG but also ToC, where the B side and G side ultimately lead back to the C side. Products like Tencent Meeting and WeCom have followed this path. They did not rebuild another Tencent but enabled Tencent to learn how to develop cloud-based software and to stop monetizing directly through user traffic.

On its 930th anniversary, Yang Guoan said that Tencent is starting to break away from the "second-generation rich" reliance on social media and gaming. This statement is only half correct; the path is indeed changing, but the dependence remains. Many years later, Tang Daosheng would say that large models, like the cloud, are hard to monopolize. The confidence behind this statement stems from the five years of reducing losses.

In May 2023, Liu Chiping stepped down as an executive director. He remains the president and still oversees the Investment Committee. However, the symbolic meaning is clear. A person who advocates that "capital is a core competency" has stepped down from the board. The media calculated that his total remuneration over 16 years exceeded 2.3 billion yuan.

From 930 to 2023, Tencent spent almost five years primarily "dismantling." Dismantling organizations, dismantling investments, dismantling visions, and dismantling horse races. It recognized its illness and therefore took off its old clothes one by one. However, the most challenging part for a company is not changing direction but acknowledging that even after changing direction, it still does not have the answer. This second transformation to the Industrial Internet ultimately ended with a reduction in losses.

Changing the course of a river, water is still water.

Tide

On May 19, 2023, at a shareholder meeting in Hong Kong, Ma Huateng was asked about AI. He said:

"AI is a once-in-a-century opportunity for the Internet."

He added that Tencent would not release half-baked products to boost its stock price.

This statement contains both clarity and tardiness. Clarity in the sense that he can see the magnitude of this revolution. Tardiness in the sense that when he made this statement, Baidu's Wenxin Yiyuan model had been released two months ago, Alibaba's Tongyi Qianwen model had been released a month ago, and Tencent's model still had to wait another four months. On September 7, 2023, the Hunyuan Mega Model was officially released, focusing on "eliminating nonsensical speech."

Security and restraint, Tencent's consistent stance.

After nearly a year of restraint, in November 2024, Tencent open-sourced the Hunyuan-Large model. With 389 billion parameters, it was the largest open-source MoE model globally at the time. As a company that has traditionally been reserved about open-sourcing, this was the first time they opened the doors to their model.

On May 30, 2024, Tencent Yuan Bao was launched. The company's public statement was: "Great models do not compete for the moment."

A year later, it contradicted itself.

On February 13, 2025, Yuan Bao integrated into DeepSeek-R1 Full Blood Edition for free. Three days later, WeChat Search launched AI Search in grayscale, also integrating DeepSeek. Tencent specifically stated that this feature would not read user chat records or Moments. On March 4, Yuan Bao topped the Apple App Store's China region free download chart.

This was the last time the old traffic magic spell manifested. By integrating a competitor's model and infusing it with its own traffic, it reached the top in a month.

Then, Tencent wanted to perform once again.

At an employee meeting on January 26, 2026, Pony Ma announced a play: "Yuan Bao Plan," distributing 1 billion red envelopes during Chinese New Year. He said he hoped to recreate the moment of WeChat red envelopes eleven years ago.

On Chinese New Year's Eve in 2014, WeChat Red Envelopes were launched. With 8 million users and 40 million red envelopes, Jack Ma called it "the attack on Pearl Harbor." One year later on Chinese New Year's Eve, WeChat collaborated with CCTV's Spring Festival Gala, prompting people nationwide to shake their phones at the TV. Overnight, red envelopes were sent and received 1.01 billion times. The landscape of mobile payments was rewritten between these two New Year's Eves.

That was Tencent's last surprise attack achieved through products and traffic. In the eleven years that followed, it never had another night like that.

During the 2026 Chinese New Year, Tencent attempted to reclaim that night with 1 billion red envelopes.

It did not succeed. Shortly after the red envelope event went live, WeChat made its move. Around February 4, WeChat restricted Yuan Bao's red envelope link opening, citing "guiding users to frequently share links" and "causing harassment to users."

Their own product was blocked outside by their own access control.

Yuan Bao responded, stating that they were urgently adjusting the sharing mechanism. WeChat's PR director gave a four-word response:

"Same treatment for all."

To be fair, WeChat was not entirely indifferent. Invitations from the Yuan Bao Plan were initially allowed to spread through WeChat friends, group chats, and Moments. This was Tencent's most essential relationship entry point, and WeChat opened a small crack. However, Yuan Bao ultimately could not squeeze through that crack.

At its peak, Yuan Bao had over 50 million daily active users and 114 million monthly active users. Then, the tide receded. Yuan Bao fell out of the top ten in the Apple Store's free list. Subsequent data from QuestMobile showed that once the red envelope event ended, Yuan Bao's daily active users returned to the pre-Chinese New Year level.

By March 2026, the monthly active users of Yuanbao's standalone app reached 57.35 million, with only three domestic AI applications having over 100 million monthly active users: Dou Bao, Qian Wen, and DeepSeek.

Three months later, during the shareholders' meeting on May 13, 2026, Ma Huateng uttered a heartfelt statement:

“A year ago, we thought we had boarded the ship, only to realize that the ship was leaking. We started to switch to another ship, and now we feel like we have stepped aboard, but cannot sit down yet. We still hope the ship can speed up.”

In 2014, WeChat's red envelope changed the payment landscape overnight during Chinese New Year's Eve. In 2026, Tencent spent 10 billion to buy back the same lesson. Traffic is not users, water is still water—it can catapult anything to the top, yet it cannot retain anything.

In June 2026, Tencent made a 100 billion move by leading the first round of funding for DeepSeek. Some interpreted it as Tencent admitting that they couldn't do it themselves and had to pay for it instead.

A 27-year-old Chief Scientist cannot change the inertia of a hundred-thousand-strong company. The lessons learned by the AI Lab over nine years are there. It's not a lack of talent; it's that this company used to, in its own best way, digest talent and dreams together.

Countercurrent

In 2026, counterflow.

Let's start with money. In the full year of 2024, Tencent's capital expenditure was 76.7 billion yuan, a 221% year-on-year increase. In 2025, it was 79.2 billion. In the most intense year of the AI arms race, this number only rose by 3%. Meanwhile, ByteDance's capital expenditure during the same period is estimated at over 160 billion, and Alibaba announced a three-year investment of 380 billion.

Tencent used the money it didn't spend on cards for repurchases.

In 2025, it repurchased about 80 billion Hong Kong dollars' worth of shares, and with dividends, the total amount returned to shareholders exceeded 240 billion Hong Kong dollars in two years. During the annual report conference call on March 18, Liu Chiping explained that in 2025, due to GPU supply constraints, they couldn't buy cards. In fact, it's not that they couldn't buy, but rather they didn’t want to buy at a high premium. The day after the financial report, Tencent's stock price dropped by over 6%.

Then, in 2026, things changed. In the first quarter, capital expenditure was 31.9 billion. In the second quarter, 52.78 billion. The total for the first half of the year was 84.68 billion, surpassing the full year of 2024. During the conference call, Liu Chiping even hinted that they might reduce repurchases in the future and invest the money into AI. The abacus that had been calculating ROI for ten years was finally set aside in 2026.

Let's first look at the organization. In December 2025, Tencent upgraded its large-scale model development architecture and established the AI Infra Department, AI Data Department, and Data Computing Platform Department. In March 2026, the AI Lab was dissolved. The carrier of the first dream was dismantled, and personnel and assets were integrated into Hybris. A company dismantled its own nine-year-old research institution, holding a funeral and handover ceremony on the same day.

Next is the individual. In December 2025, Tencent announced that 27-year-old Yao Shunyu was appointed Chief AI Scientist of the "CEO/President's Office" and would report to President Liu Chiping.

Yao Shunyu, born in 1998, is from Hefei, Anhui. He attended Hefei No. 1 High School, ranked third in the province in the 2014 college entrance examination for science, and was admitted to Tsinghua University's Yao Class. At Tsinghua, he organized a rap society. His high school homeroom teacher remembers this child who ranked third in the province in science, liked to rap, dance, and play basketball.

He pursued his Ph.D. at Princeton University, during which he developed the ReAct framework and the Thought Tree, and led the SWE-bench project. Today, half of the world's intelligent body of research is based on the scaffolding he built. In August 2024, he joined OpenAI, contributing to Operator and Deep Research. He left in September 2025. Three months later, he became Tencent's Chief AI Scientist at the age of 27.

When asked why he chose Tencent, his answer was:

"The most important thing is everyone is very sincere."

On January 10, 2026, after joining, he made his first public statement, talking about how "Claude Code is reshaping the computer industry." On June 5, he had a conversation with Tang Daosheng and was asked if Tencent AI was lagging behind. He said:

"AI is a long-term game, and the second half is just beginning."

Tang Daosheng added: "This is a long-distance race, this is a marathon, and Tencent still has very rich scenarios."

In Tencent's AI landscape, there are two armies.

One is in TEG, Hybris. After Yao Shunyu took over, it consolidated the model development, data, and training infrastructure scattered throughout the organization. The dissolved AI Lab saw some research personnel integrated into the Large Language Model Department, while others moved to the Industry-University-Research Cooperation Center.

The other is in WXG, WeChat. Its name is WeLM. In 2022, WeChat released this Chinese pre-training model with 10 billion parameters, surpassing similar models on 18 Chinese tasks. In August 2025, WeChat unveiled WeChat-YATT, a large-scale model reinforcement learning training framework. By 2026, the new generation of WeLM shifted to a sparse MoE architecture, with a base version of 800 billion parameters activating only 30 billion at a time; the extended version has 1.3 trillion parameters, activating 49 billion. On July 9, the WeChat AI team publicly shared the Hidden Decoding research, detailing two new models in a paper.

People are also moving back and forth. By the end of 2025, Xu Can, a senior researcher at MixNet, transferred to WeLM. Xu Can is the creator of WizardLM and proposed Evol-Instruct at Microsoft in 2023, allowing large models to generate challenges for themselves. After he left, MixNet's post-training team split into several small groups, each forging ahead on their own.

Two teams, two different logics. MixNet holds centralized computing power, data, and model resources. WeLM has WeChat's native entry point, mini-program services, and real feedback from over a billion users. They collaborate and compete. Tencent's AI entry-point battle happened right at its own doorstep.

In 2019, Zhang Xiaolong said, "Every day, one hundred million people teach me how to create products." On January 9, 2020, during a WeChat open class, Zhang Xiaolong was absent. He recorded a 13-minute video discussing seven considerations on information interconnection, admitting two mistakes on the public platform, and announcing the intention to create short-form content.

A little over ten days later, Video Number was launched. This was WeChat’s only direct response during the fierce competition with Douyin in those years. That battle later proved one thing: Tencent's product capability was not dead; it had always resided within the walls of WeChat.

In 2021, he summarized WeChat's ten years in two words: connection and simplicity.

He said he hoped that WeChat would always be like a small and beautiful product, with its own soul, aesthetics, creativity, and ideas. Back in the day, Pan Luan used to call him the "fool among geniuses."

During the 2026 Chinese New Year, it was this "fool's" obsession with product perfection that blocked Ant Group's 1 billion red envelopes from entering WeChat. The most restrained part of Tencent turned out to be the most like a product company. In 2026, WeChat was put on the spot. Search integrated with AI, WeChat Agent began testing, and the AI assistant "Xiao Wei" was quietly released. Some said that this time, Zhang Xiaolong had wagered everything.

Large models had long since entered WeChat's ecosystem. In the ad recall stage of Video Number and Moments, there was a system called LEADRE, which deployed a MixNet model with 1 billion parameters at its core.

In June 2026, "Xiao Wei" began a small-scale grayscale test. Users could directly command it to operate WeChat's native functions and call mini-programs to complete tasks. Insiders who participated in the testing said that WeLM served as the main model for Xiao Wei, and some responses would invoke DeepSeek.

In June, the WeChat Open Platform did something more comprehensive. Mini-programs could now integrate with the WeChat AI ecosystem. In automatic mode, the platform would read the mini-program's source code, analyze the page structure, and allow WeChat AI to operate directly. In development mode, developers encapsulate business capabilities into skills, which, once approved, are available for WeChat AI to call upon. WeChat aimed to nurture AI to be an extension of itself.

Lastly, there was that 100 billion deal. In June 2026, DeepSeek completed its first round of financing, totaling over 50 billion dollars. Liang Wenfeng personally put in around 20 billion, CATL around 5 billion, and Tencent around 10 billion. It is said that the investors were personally selected by Liang Wenfeng.

Investment is Tencent's best posture. However, this investment can also be seen as a new attitude. Eight years ago, Tencent's investment was to "constrain Alibaba," a defensive move. Today, it is investing in a company that it currently cannot catch up with, and has even opened up its most valuable traffic channel, WeChat, to its model. The kind of company that used to be "acquired" by Tencent's investment is now recognized as a teacher by it.

Water flows downhill by nature; swimming against the current is the business of those who swim upstream.

In 2026, Tencent, while continuing to flow, has not lost its flow, investment, or ecosystem. At the same time, it is starting to swim upstream, pouring in real money and assets into uncertainty without seeing a return. Its dream has returned.

Epilogue

The conclusion of "Tencent Has No Dream" reads:

"Hopefully Tencent won't have to wait another four and a half years to review and reflect."

Tencent has waited not just four and a half years; it has been eight years.

During these eight years, it distributed "capital" to shareholders, opened up "traffic" to competitors, and sent the winner-takes-all mechanism into history. The old vision has been replaced with a new one, and the AI Lab it built with its own hands has been dismantled. Finally, a 27-year-old young man took up the position of Chief Scientist. It has experienced all the falls that were once foretold— the decline of the ingots, the reduction in cloud losses, and the funeral of the first AI dream. After each fall, it still moves in that direction.

"The Great Gatsby" was written in 1925. Gatsby was a man who staked his entire life on the "past." He bought a mansion on the opposite shore of Long Island Sound, hosting night after night of parties, all lit up, just for a green light across the bay. Behind that light lived the girl he loved five years ago. In the final pages of the novel, Fitzgerald wrote a passage for Gatsby, and for all those who want to rewrite time:

"So we beat on, boats against the current, borne back ceaselessly into the past."

Gatsby swam upstream to return to the past. He believed that as long as he paddled fast enough, time would reverse, and old dreams would reignite. He failed. Bullets pierced the pool, the party ended, and that green light remained across the bay.

Tencent is also rowing a boat, but it is rowing against the current to never return. To never go back to the self that only cared about capital and traffic, only about certainty, living a life as fluid as water.

Tencent still has dreams.

Original Article Link

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia