Summary of Key Points
Tencent's latest financial report revealed a negative free cash flow of 13.8 billion yuan—this is the first time the company, known as the "king of profit cash," has experienced such a situation. The reason behind this is the massive capital investment required for its AI transformation. The core argument of the article is that AI is not a low-cost upgrade of the internet industry; rather, it represents a high-capital model that requires substantial investment in building computing power infrastructure. Giants like AT are engaging in an "arms race" driven by fear: they fear being left behind by the times and thus feel compelled to invest, yet the path to profitability is unclear. As a result, the market is re-evaluating Tencent using manufacturing industry standards, leading to a significant drop in its stock price. This transformation is essentially a dangerous game with no clear end in sight.
1. AI Is Not an "Upgraded Version" of the Internet, but a "High-Capital Digital Factory"
In the past, the internet industry made money through low-cost models: for example, Tencent's WeChat could serve one billion users with a single development effort, generating cash flow with almost no additional cost. It was assumed that large AI models (such as Tencent's Yuanbao and Alibaba's Qianwen) would continue this pattern of low-cost profitability. However, the reality is quite different: AI models are like factories that produce tokens (AI-generated content). These models require significant upfront investment in purchasing GPU servers and building data centers (fixed costs), and they consume computing power and electricity with each user query (variable costs). Only when a sufficient number of users are engaged (high capacity utilization) can the model be profitable. This is similar to operating a factory—once the equipment is purchased, it depreciates, and the higher the production volume, the lower the cost per unit. Therefore, AI is a high-capital endeavor that has nothing in common with the low-cost nature of the internet industry.
2. Negative Free Cash Flow: AI Is Eating Away Giants' Profits
Tencent's core businesses (games, advertising) are growing well, but the money invested in AI exceeds the profits generated:
- Alibaba's net profit declined by 99% last quarter due to 38.7 billion yuan in AI-related expenses.
- Tencent has already invested over 150 billion yuan in AI and continues to spend more, resulting in a negative cash flow.
This situation is similar to the real estate industry: developers must invest in new projects before they have recouped the costs from previous ones, often relying on borrowing. Although NVIDIA's performance appears impressive, it too relies on financing from Wall Street. The AI industry chain is currently driven by upstream players, with no support from end-users; growth comes from internal funding rather than real demand.
3. Stock Price Drops by 5%: The Market Is Re-Evaluating Tencent Using Manufacturing Standards
Tencent was previously valued highly because it could generate stable free cash flow, which could be used for dividends and share repurchases. However, the market's perspective has changed:
- Internet companies are valued based on their free cash flow, while manufacturing firms are evaluated based on capacity utilization and payback periods.
Tencent's AI investment is driven by fear of being outpaced by competitors. There is no clear return on investment, and the market dislikes this endless arms race, leading to a lower valuation using manufacturing industry standards and a consequent drop in its stock price.
4. The Road to Profitability Is Difficult: Tencent's "Backup Plan" Lacks Clarity
Tencent claims that it could still profit by renting out its computing power even without developing AI applications. However, the reality is bleak:
- It uses most of its computing resources for internal models (e.g., Yuanbao), and consumer willingness to pay for these services is very low—even Americans are reluctant to pay.
- Lacking key customers: Tencent's internet ecosystem does not have large enterprises like Meituan or JD.com that would rent computing power, unlike foreign cloud providers (e.g., Google and Microsoft) with lucrative contracts (e.g., $500 billion in orders generating 30% profit).
Therefore, the idea of profiting from renting out computing power remains more theoretical and difficult to replicate on a global scale.
5. The AI Arms Race Resembles the Moon Race Between the US and the USSR: A Dilemma with No Clear Solution
This investment in AI is essentially based on fear-based decision-making: all parties fear being left behind, similar to the Prisoner's Dilemma—if you don't invest, your competitors will; if you do, you can't see the end of the investment. It's like the US and USSR's moon race: neither country aimed to explore the moon but to outdo the other. In the AI industry, there is much talk about AGI (Artificial General Intelligence) as the future, but have Tencent and Alibaba's leaders truly understood what this means? Perhaps not. However, no one dares to stop investing for fear of being eliminated.
The only possible solution would be for a dominant player to establish rules (e.g., limiting excessive investment), but for now, this dangerous game continues.
Conclusion
AI is a future trend, but the transformation efforts by giants are akin to a gamble where they risk their profits to secure their survival. The market dislikes such uncertain investments, which is why Tencent's financial performance remains stable yet its stock price has plummeted. The outcome of this situation will only become clear when AI becomes commercially viable (e.g., when users are willing to pay and cloud orders surge). For now, all parties are pushing forward with little clarity about the future.