Summary of the Key Points in Plain Language
This analysis of the AI valuations of 30 established Chinese and American tech companies directly challenges the common misconception that in the AI era, capital prefers new companies over established ones. Capital doesn’t care about a company’s “age” and won’t casually invest for the long term in just any “AI story.” All valuation considerations boil down to two practical questions: Are your assets essential for the current AI industry? And after deducting all AI-related costs, can you actually generate substantial profits?
Explanation of the Five Dimensions in Simple Terms
1. Why does Baidu, with over half of its revenue from AI, have a smaller market value than a fraction of Dell’s?
Many wonder why Baidu, which started AI earlier than most domestic companies and now generates more than half of its revenue from AI, and SenseTime, whose generative AI business has surpassed its traditional businesses and even turned a profit, still has a lower valuation. The reason is that the market doesn’t allow for separate evaluations of old and new businesses. You can’t pretend that the burdens from past decades simply disappear. Baidu’s traditional search advertising business is still declining, and its AI business continues to spend billions on computing power and cloud services. These costs and risks are included in the same financial reports. It’s like running a restaurant where a new, popular dish generates half of the revenue, but the old business is still losing money. Investors won’t treat Baidu as a completely new, profitable brand; they’ll value it based on its mixed performance.
2. Have Dell and Sandisk, with no new stories to tell, seen their stock prices rise fivefold in two years? Is that just luck?
Contrary to Baidu and SenseTime, these once-popular hardware companies (Dell, Sandisk, Cisco) have seen their stock prices soar despite having outdated products. This isn’t luck but the result of the “scarcity migration” in the AI industry. Everyone focuses on expensive GPUs, but AI relies on physical infrastructure. Even with thousands of high-performance GPUs, if the servers can’t handle the data or the power supply and cooling systems fail, the GPUs are useless. New companies can’t quickly build the necessary infrastructure; it takes years to expand production. Capital flows to where the industry’s growth is most constrained, regardless of the company’s age. These established companies have built strong supply chains and engineering capabilities that are essential for AI development.
3. Leveraging AI trends to boost stock prices only lasts for a few quarters.
Many companies add a few lines about AI in their annual reports, and their stock prices immediately rise. However, this “AI story premium” doesn’t last long. Investors are willing to invest in future potential, but only for a short period. For example, if a company claims 44% of its PC shipments have AI features or if AI usage has increased by 256%, it can temporarily boost its stock price. But if it can’t show that AI actually increases sales or profits, the premium will vanish. Examples include HP and Zoom, whose stock prices rose despite little actual AI-related revenue.
4. Why are AI subsidiaries more valued than their parent companies?
Some companies’ AI subsidiaries are valued higher than their parent companies. This is because valuable AI assets within mature companies are often underutilized. To realize their full value, these assets need to be spun off or integrated into existing businesses. If they are spun off, new investors will evaluate them separately, but the parent company’s shares will be diluted, reducing its overall value. If they are integrated, the parent company can use AI to improve its existing businesses, but the high capital costs (e.g., billions for computing power) can eat into profits.
5. Why does capital invest in established companies?
The key factor for capital investment is whether a company is a crucial link in the AI value chain. Even if a company is old, if its assets are essential for AI development and can drive growth, it can still be valuable. For example, if an old company’s infrastructure and customer base complement AI, it can thrive. However, if its old businesses hinder growth, capital won’t invest. The ultimate question is whether the company can generate stable profits from AI.
In summary, capital doesn’t care about a company’s age; it only cares whether it can generate sustainable profits from AI. Established companies with the right assets and capabilities can still be valuable in the AI era.