Thinking about Zhipu AI now, it’s also a non-consensus investment opportunity. Its value has nearly doubled in a week, with a market cap now under 100 billion.

When it went public, everyone thought its business model was to G (Government), and a large part of the company's operating profit came from providing AI integrated machines(private deployment) for enterprises and governments. They even hired many salespeople for this, which is usually a one-time purchase.

But what the market overlooked is that GLM4.7 is truly powerful. In fact, as early as two weeks ago, Zhipu was running out of selling capacity due to high sales, and had to impose daily purchase limits. Today, with the release of GLM5, the product has been fully priced higher, roughly doubling the previous price. The only standard for evaluating a company is whether it has the ability to raise prices. However, AI changes so quickly that models which are leading today are often surpassed in just a few months. DeepSeek has only been around for a year so far.

Also, the current AI market cap is actually quite expensive. It reminds me of the previous boom in new energy vehicles. Now, with BYD, Xiaopeng, and NIO, the highest market cap is Li Auto at only 150 billion. Xpeng and NIO are even less likely to return to their previous highs.

Finally, the competition among AI companies might be similar to the previous competition in new energy vehicles. The entire market track has expanded, profits have increased, and business models have been proven, but stock prices are still not as high as before the companies were profitable. However, optimistically, AI companies still have a chance to reach 10% of Tencent’s market cap. Although Tencent and Alibaba seem less prominent than these trending AI companies, they have already invested early. Alibaba holds over 30% in Yuezhiyin, over 10% in Minimax, and over 5% in Zhipu. Tencent also has stakes in these companies, roughly between 2.5% and 5%.