Alibaba's CEO has declared a new mission to industrialize artificial intelligence.
Wu Yongming recently told investors that the company's servers are running at near-full capacity, a clear sign of the immense demand for AI processing power. To capitalize on this, he introduced a strategy centered on building two types of 'core factories': one for AI training and another for AI inference. This isn't just about adding more hardware; it's a fundamental plan to scale up their computing supply, increase prices to improve profitability, and turn every bit of AI usage into a steady stream of revenue.
So, what led to this 'AI factory' announcement? The context is built on several key recent events.
First, the announcement was strategically timed with Alibaba's latest earnings report. While overall revenue growth was a modest 3%, the company highlighted its cloud and AI divisions as key drivers. The 'two factories' narrative provides investors with a clear roadmap for how Alibaba plans to invest its capital and adjust pricing to boost margins in these promising sectors.
Second, Alibaba has already laid the groundwork for monetization. In April, Alibaba Cloud raised prices on its AI computing and storage services by as much as 34%. This move, which the company attributed to rising hardware costs, directly supports the idea that its resources are in high demand and reinforces its ability to command higher prices—a core principle of the factory model.
Third, the company is actively building out its own infrastructure to reduce reliance on foreign technology. In a significant step, Alibaba, in partnership with China Telecom, recently launched a massive 10,000-card computing cluster powered by its in-house 'Zhenwu' accelerator chips. This demonstrates a tangible commitment to constructing these AI factories using domestic silicon, a crucial move given the geopolitical climate.
Finally, this strategy is also a response to significant external pressures. The United States has been tightening its restrictions on advanced chip exports to China. Recent reports even alleged that high-end Nvidia servers were smuggled into the country, with Alibaba named as a potential recipient. This regulatory scrutiny makes building a self-sufficient, domestic supply chain not just a business advantage, but a strategic necessity.
In short, Alibaba's 'AI factory' strategy is a comprehensive response to soaring demand, a tool for margin recovery, and a defensive measure against international supply chain risks.
- AI Training: The process of 'teaching' an AI model by feeding it enormous datasets. This is computationally intensive and is the foundation for creating powerful AI.
- AI Inference: The process where a trained AI model applies its knowledge to make predictions or generate responses to new, unseen data. This is what happens when you use an AI application.
- Tokens: The fundamental units of data (like words or parts of words) that AI models process. The cost of using AI services is often calculated based on the number of tokens used.