On September 24, Alibaba's stock price hit a four-year high, rising more than 9% in a single day, and its market value reached approximately HK$3.32 trillion, an increase of nearly HK$300 billion in one day.

What triggered the stock price surge was Alibaba Group's annual Yunqi Conference. At the conference, Alibaba released several major announcements, and especially after the speech by Wu Yongming, CEO of Alibaba Group and Chairman and CEO of Alibaba Cloud Intelligence Group, Alibaba's stock price showed a significant upward trend.

图片

At this grand event in the field of AI and cloud services, Wu Yongming shared that achieving general artificial intelligence (AGI) is a certainty, but this is only the starting point. The ultimate goal is to develop super artificial intelligence (ASI) that can self-iterate and completely surpass humans.

"Alibaba Cloud will continue to increase investment to welcome the arrival of the era of super artificial intelligence."

In fact, since clearly defining its "user-first, AI-driven" strategy in 2023, Alibaba has consistently focused on its two main businesses, "e-commerce and AI + cloud," and has made unwavering investments. At the beginning of this year, Alibaba once again proclaimed its "All in AI" slogan. With continuous and intensive investment, Alibaba has formed a full-stack technical capability encompassing AI computing power, AI cloud platform, AI models, open-source ecosystem, and AI applications.

"The industry's demand for AI infrastructure far exceeds our expectations," Wu Yongming revealed. Alibaba is advancing a 380 billion yuan AI infrastructure construction project and plans to make even greater investments.

Alibaba Cloud will strive to become a full-stack AI service provider. To prepare for the arrival of the ASI era, compared to 2022, the energy consumption of Alibaba Cloud's global data centers will increase tenfold by 2032. This means that Alibaba Cloud's investment in computing power will increase exponentially.

It's not hard to see that Alibaba's strategic plan on the road to ASI is already very clear.

图片

A clear path toward ASI

The Yunqi Conference is not only a grand technology event, but also a trendsetter in the fields of AI and cloud computing. It has witnessed every wave of China's technological development and provided clues for understanding the future development of the industry.

At this year's Yunqi Conference, themed "Cloud-Intelligence Integration and Carbon-Silicon Symbiosis," Wu Yongming also systematically elaborated for the first time on the three-stage evolution path to ASI:

The first stage: "Intelligence Emergence", where AI acquires generalized intelligence by learning from massive amounts of human knowledge.

The second stage: "Autonomous Action," where AI masters the ability to use tools and program to "assist humans," is the current stage of the industry.

The third stage: "Self-iteration," where AI connects to the physical world and learns on its own, ultimately "surpassing humans."

图片

Wu Yongming pointed out that achieving the second stage of the leap mainly lies in breakthroughs in tool use and coding capabilities. AI will rapidly penetrate all industries, including manufacturing, logistics, finance, and scientific research, as a powerful productivity tool. In the future, there may be more agents and robots than humans working alongside them, having a huge impact on the real world.

In the third stage, the final hurdle to overcome for ASI, AI not only needs access to all raw data from the real world, but also needs the ability to self-learn. By continuously interacting with the real world to obtain feedback, and with the help of reinforcement learning mechanisms, AI will be able to achieve self-correction and intelligent upgrades, ultimately evolving into ASI that surpasses human intelligence.

In terms of specific operations, Alibaba has also provided a clear strategic path.

According to Wu Yongming, Alibaba Cloud, as a "full-stack artificial intelligence service provider," will implement the ASI strategy through two core paths: First, Tongyi Qianwen firmly adheres to the open source and open approach, and is committed to creating "the Android of the AI ​​era"; second, it builds a super AI cloud as a "next-generation computer" to provide intelligent computing power networks for the world.

Judging from the stock price performance, the capital market clearly approves of Alibaba's AI vision. Astute investors have been betting on this internet giant for some time, including Wall Street star fund manager Cathie Wood.

According to the daily trading report of Ark Investment, managed by Wood Sister, the company's two ETFs—ARKF and ARKW—purchased Alibaba's American Depositary Receipts (ADRs) on September 22, totaling approximately $16.3 million. This marks the first time the company has re-established a position in Alibaba since 2021, and it is also the first time Wood Sister has held shares in the company in nearly four years.

In the second quarter of this year, Scion Asset Management, the firm of renowned hedge fund manager Michael Burry, also sold put options on Alibaba and bought call options on it. In July, Invseco Developing Markets, the flagship emerging markets fund of US asset management giant Invesco, also increased its holdings in Alibaba.

图片

Positive cycle brought about by cloud services

Focusing on large-scale models and AI cloud as key levers on the path to ASI, Alibaba also launched more than ten new products at the Yunqi Conference.

Among large language models, Alibaba's flagship model, Qwen3-Max, has made its debut, outperforming GPT5 and Claude Opus 4, ranking among the top three globally. Qwen3-Max includes two versions: Instruct and Thinking. Its preview version has already ranked third on the Chatbot Arena leaderboard, and the official version is expected to achieve further performance breakthroughs.

图片

Tongyi's large-scale models have become the world's number one open-source model and the most chosen model by Chinese enterprises. To date, Alibaba's Tongyi has open-sourced over 300 models, covering "full-size" models of varying sizes and "full-modality" models including LLM, programming, image, speech, and video. Global downloads have exceeded 600 million, with 170,000 derivative models worldwide, firmly holding the top position globally. Over 1 million customers have integrated Tongyi's large-scale models. According to Frost & Sullivan's 2025 H1 report, Alibaba's Tongyi holds the largest market share in China's enterprise-level large-scale model invocation market.

In terms of AI infrastructure, Alibaba Cloud has carried out collaborative optimization and system innovation across the entire software and hardware stack around AI, and has initially formed an operating system centered on Tongyi and a next-generation computer centered on AI cloud. In the past year, Alibaba Cloud's AI computing power has increased more than fivefold, and its AI storage capacity has increased more than fourfold.

The Computing Pavilion at this year's Yunqi Conference showcased a number of comprehensively upgraded Alibaba Cloud AI infrastructures, demonstrating Alibaba Cloud's full-stack AI technology capabilities, from underlying chips, supernode servers, high-performance networks, distributed storage, intelligent computing clusters to artificial intelligence platforms and model training and inference services. In addition, Alibaba Cloud also exhibited for the first time a high-density AI server supporting 144 computing nodes per rack and its next-generation high-performance network architecture, HPN8.0.

On the other hand, Alibaba's investment in AI infrastructure has already yielded tangible revenue returns, moving into a positive cycle of high investment and high growth.

According to Alibaba Group's financial report, the Group's AI+cloud business performed strongly in the first quarter of fiscal year 2026 (the second quarter of calendar year 2025), with Alibaba Cloud's revenue growth continuing to accelerate to 26%, a three-year high. Driven by eight consecutive quarters of triple-digit growth in AI-related revenue, AI revenue accounted for more than 20% of external commercial revenue in this quarter.

In terms of market share, Alibaba Cloud has also demonstrated a significant competitive advantage.

On September 9th, international research firm Omdia released a report titled "China AI Cloud Market, 1H25". The report predicts that in the first half of 2025, the Chinese AI cloud market will reach 22.3 billion yuan, with Alibaba Cloud leading the pack with a 35.8% market share, exceeding the combined market share of the second to fourth ranked companies. The report also forecasts continued explosive growth in the Chinese AI cloud market over the next few years, with sales expected to reach 51.8 billion yuan in 2025 and potentially 193 billion yuan by 2030.

图片 图片

Deep Links to "Carbon-Silicon Symbiosis"

From cutting-edge technologies to real-world scenarios, AI has already penetrated into the details of human life with unprecedented depth and breadth.

This change is most noticeable in the cutting-edge application pavilion of the Yunqi Conference. Hundreds of robots of various shapes and sizes are silently "breathing" in the venue. They are not only reliable enough to steadily lift toolboxes, turn around and raise their arms, but also elegant enough to make latte art coffee, and cute enough to provide emotional value by providing companionship to humans.

"The more powerful AI becomes, the more powerful humanity becomes."

Regarding the question, "After the emergence of ASI, where AI becomes increasingly powerful and may even surpass human intelligence, how will humans and AI coexist?" Alibaba has provided an answer: "We are optimistic about the future. With the arrival of super artificial intelligence, humans and AI will have a brand-new way of collaborating."

At the same time, Alibaba Cloud also announced at the Yunqi Conference that it will collaborate with NVIDIA in the field of Physical AI. This collaboration will provide enterprise users with a full-chain platform service, including data preprocessing, simulation data generation, model training and evaluation, robot reinforcement learning, and simulation testing, further shortening the development cycle of applications such as embodied intelligence and assisted driving.

图片

Physical AI is a branch of AI technology that extends into the physical world. It aims to enable AI to interact with the real world by integrating multimodal perception, spatial relationship understanding, and physical rule cognition. Typical applications include humanoid robots, advanced autonomous driving systems, and other systems that require interaction with the physical environment.

According to reports, Alibaba Cloud's AI platform PAI will integrate the complete Physical AI software stack, including NVIDIA Isaac Sim, NVIDIA Isaac Lab, NVIDIA Cosmos, and Physical AI datasets. Combined with the capabilities of Alibaba Cloud's big data AI platform, it will form a full-chain platform support covering data preprocessing, simulation data generation, model training and evaluation, robot reinforcement learning, and simulation testing.

It's clear that this collaboration is not simply a combination of technologies, but a deep integration covering the entire interaction chain between AI and the physical world. For Alibaba, with its massive business, the further interaction between AI technology and the physical world will impact far more than just Alibaba Cloud; it may also bring about significant changes to businesses more closely integrated with human life, such as e-commerce and local services.

From the grand ASI blueprint to the continued upgrading of AI infrastructure, and further exploration of how AI technology can deeply connect with the physical world, Alibaba presented a clear, progressive, and already-delivering AI path at the 2025 Yunqi Conference.

However, as Wu Yongming said, the speed of AI development has surpassed any historical period, but we are only in the early stages of AGI transformation. The greatest imagination of generative AI is not to create one or two new super apps on mobile phone screens, but to take over the digital world and change the physical world.

Therefore, "carbon-based symbiosis" should not be seen as a distant end point; rather, it is a technology-driven reality unfolding before our eyes. Its occurrence is evident in every moment of collaborative decision-making between humans and AI, and embedded in every operation of intelligent systems that optimizes the real world.

图片

With the arrival of the ASI era, it remains to be seen what kind of multiplier effect Alibaba, as the only company in China's cloud computing field with full-stack AI capabilities, will demonstrate.