Even as light rain fell over Hangzhou, the atmosphere at Yunqi Town buzzed with excitement. On September 24, Alibaba Cloud opened its annual Apsara Conference 2025, where CEO Eddie Wu shared a powerful vision for the future of artificial intelligence — one that moves beyond today’s smart tools toward what he called artificial superintelligence (ASI).
From Ideas to Action
A year ago, when Wu first took the stage as CEO, he spoke about AI’s potential to transform both the digital and physical worlds. This year, those ideas became concrete plans. Alibaba Cloud announced major new AI models, massive infrastructure investments, and a clear roadmap toward ASI — a level of intelligence that can eventually surpass human capability.
A Wave of New AI Releases
Alibaba Cloud introduced a full lineup of new AI products and models, led by its most powerful system yet — Qwen3-Max.
- Qwen3-Max: A flagship AI model with more than a trillion parameters, trained on 36 trillion data points. It reportedly outperforms leading systems like GPT-5 and Claude Opus 4 in global benchmarks.
- Qwen3-Next: A cost-efficient model that uses a new architecture to cut training expenses by 90%.
- Qwen3-VL: A “vision” model that can interpret charts, images, and even turn design sketches into working web code.
- Qwen3-Coder: An improved programming model that writes cleaner, safer, and faster code.
- Qwen3-Omni: A multimodal model that understands text, audio, and video together.
- Wan2.5-Preview: A model for text-to-video and video editing.
- Fun: A family of speech models with lifelike voices for e-commerce, audiobooks, and entertainment.
These launches show Alibaba Cloud’s determination to lead the global AI race not just with one model but with an entire ecosystem.
Redefining the Future of Computing
In his keynote, Wu shared two bold ideas shaping Alibaba Cloud’s strategy:
1. Large AI Models Are the New Operating System
Wu believes that in the near future, traditional software will fade into the background. Instead of opening separate apps, people will simply talk to intelligent agents — powered by large models like Qwen — to get things done.
In other words, AI will become the interface for everything, from writing emails to managing businesses.
2. The “Super AI Cloud” Is the Next Computer
Wu compared today’s shift to previous computing revolutions — from mainframes to PCs to smartphones. In the AI era, he said:
- Natural language is the new programming language.
- Agents are the new software.
- Context is the new memory.
- Large models act as the “operating system” that connects people, data, and machines.
To make this happen, Alibaba Cloud is building what Wu calls a “super AI cloud” — a global network powerful enough to train and run the most advanced models.
Massive Investment in AI Infrastructure
Earlier this year, Alibaba Cloud announced a RMB 380 billion (US$53 billion) investment plan over three years to expand its AI and cloud infrastructure. Wu confirmed that this commitment will grow even further, with a vision to increase global data center capacity tenfold by 2032.
This investment will help Alibaba Cloud deliver faster, more secure, and more scalable AI services worldwide — and prepare for the coming ASI era.
Understanding the Path to Artificial Superintelligence
Wu explained Alibaba Cloud’s roadmap to artificial superintelligence in three stages:
- Emergent Intelligence — AI learns reasoning and problem-solving from human knowledge.
- Autonomous Action — AI starts using tools and coding independently to assist humans (the stage we’re in now).
- Self-Iteration — AI connects directly to real-world data, learns autonomously, and eventually exceeds human capability.
While many companies focus on artificial general intelligence (AGI) — machines that can think like humans — Alibaba Cloud is already setting its sights beyond that, toward ASI.
The Global Race for AI Dominance
The competition to build the best AI models is fierce. After the launch of GPT-5, OpenAI faced criticism for slower innovation, while Meta and others doubled their investments. In contrast, Alibaba Cloud is moving fast and investing heavily — a signal that didn’t go unnoticed.
Following the Apsara announcements, Alibaba’s stock jumped over 9%, reaching its highest level in nearly four years.
Open Source and Community Power
Unlike some tech giants that keep their AI models closed, Alibaba Cloud has become a strong supporter of open-source AI.
- It has open-sourced over 300 models, with more than 600 million downloads and 170,000 spin-offs globally.
- The Qwen models have earned international recognition, alongside DeepSeek, as one of China’s top open-source contributions.
- Its developer platform, Model Studio, now sees 15 times more daily model requests than a year ago.
This open approach is paying off. Alibaba Cloud’s revenue grew 26% year-over-year, and its AI-related business has seen triple-digit growth for eight straight quarters.
Competing Systems, Shared Goals
Alibaba isn’t alone in its pursuit. Tencent is applying AI internally before expanding outward, while ByteDance follows a tightly controlled model similar to Apple’s iOS strategy.
Alibaba Cloud, on the other hand, has taken a “full-stack” approach — building everything from the computing hardware to the AI models themselves. This strategy mirrors Google’s, giving Alibaba end-to-end control and efficiency.
“The Android of the LLM Era”
Alibaba Cloud’s new slogan, “The Android of the LLM (Large Language Model) era,” sums up its mission: to become the open, flexible platform that powers the world’s AI applications.
Just as Android made smartphones accessible to everyone, Alibaba Cloud wants to make AI accessible to all businesses and developers.
“Tokens are the electricity of the AI world,” Wu said — meaning that AI models, like electricity, will power every digital service in the future.
The Road Ahead
AI adoption is still in its early stages. Most companies are just beginning to integrate large models into daily operations, but the momentum is clear. As demand for intelligent computing grows, Alibaba Cloud’s mix of scale, openness, and vision positions it as one of the most influential players shaping the future of global AI.
For a company that once called itself the “water and electricity of the digital world,” this new direction — as the “Android of the AI era” — represents not just evolution, but a redefinition of how the next generation will interact with technology.