After the recent round of financial report disclosures from major internet companies, there has been a quite interesting phenomenon.
Tencent fell, Alibaba fell, Baidu fell, Kuaishou fell, and NetEase also fell. Among the six companies, Xiaomi has become the rare exception.
But these companies are not without an AI story.
Baidu has Kunlun Chip and AI Cloud, Kuaishou has Kling, Alibaba has Tongyi Qianwen and Alibaba Cloud, Tencent has Hunyuan and WorkBuddy, Xiaomi has MiMo, and NetEase has already integrated AI extensively into game development.
One could even say that Chinese internet companies have never had such a strong consensus on AI as they do today.
The problem lies precisely here.
The stronger the AI consensus, the harder it is to capture an AI premium.
01
Two or three years ago, launching a large model, buying a batch of GPUs, and declaring "All in AI" were enough in themselves to construct a new valuation story. Because back then, the market was first judging: who could get a ticket to the AI era, and who would fall behind.
Now that question is basically answered.
Almost all major players have entered the AI era.
Thus, "having AI or not" quickly lost its power to differentiate, and the market began asking the next, more difficult question:
What exactly has AI changed about your company?
As soon as this question is asked, Tencent, Alibaba, Baidu, Kuaishou, Xiaomi, and NetEase immediately diverge onto six different paths.
Tencent's AI is built on top of a massive cash-generating machine.
In Q2, Tencent's domestic game revenue grew 17%, and advertising grew 22%. AI has already begun tangibly improving advertising recommendation and commercialization efficiency.
Tencent doesn't need AI to save its old businesses.
Its problem is almost the opposite: the old businesses are too good at making money, leading investors to calculate whether reinvesting so much money for AI is actually worth it.
Tencent's Q2 capital expenditure reached 52.8 billion yuan, a year-on-year increase of 176%.
This creates a contradiction for Tencent that wasn't common in the past:
On one hand, AI enhances Tencent's original business model; on the other hand, it is changing Tencent's original capital return model.
In the past, one of Tencent's most attractive features was that businesses like gaming, advertising, and social networking, once they reached scale, could generate cash flow continuously.
Now, that cash is turning back into GPUs, servers, and data centers.
So the question becomes:
How much more money needs to be spent?
02
If Tencent is building AI on its original empire, then Alibaba is trying to rebuild part of its empire around AI.
Tongyi Qianwen, Alibaba Cloud, and AI infrastructure have formed an increasingly clear chain. In the latest quarter, AI cloud and computing service revenue grew 45%, and AI product revenue achieved triple-digit growth for the 12th consecutive quarter.
This is probably one of the clearest AI commercialization paths among Chinese tech giants.
But at the same time, Alibaba's single-quarter capital expenditure also reached 67.7 billion yuan, a 75% year-on-year increase.
The more certain the AI business, the greater the investment.
This creates a very interesting turning point:
The market used to worry that Alibaba would miss AI; now it's starting to worry that Alibaba is taking AI too seriously.
After Alibaba spent recent years finally re-emphasizing profits, shareholder returns, and free cash flow, AI is now pulling the company back into a massive reinvestment cycle.
AI for Alibaba is not just a new business.
It is forcing Alibaba to re-answer a bigger question: Is it fundamentally an e-commerce company, a cloud computing company, or a tech company reorganized with AI as its underlying capability?
Baidu faces a clearly more urgent problem.
Baidu doesn't actually lack AI assets.
It has models, AI Cloud, autonomous driving, and Kunlun Chip, which is garnering increasing attention from the capital markets.
In Q2, Baidu Core AI business revenue reached 12.5 billion yuan, a 25% year-on-year increase.
Looking at AI transformation alone, this performance isn't bad.
The problem is that online marketing revenue fell 19% in the same period, and total company revenue fell 4%.
This means Baidu is facing a very brutal speed problem:
Can the speed at which the new Baidu is growing outpace the speed at which the old Baidu is shrinking?
Kunlun Chip can command a high valuation, AI Cloud can continue to grow, and autonomous driving can tell an even longer story.
But as long as search advertising, that old engine, declines faster, what the capital market ultimately sees is Baidu's overall growth.
For Tencent, AI can be an incremental addition; for Baidu, AI increasingly looks like a business replacement that must succeed.
This is also why, despite owning many assets the market considers "good," Baidu's stock still fell after the earnings report.
The value of Baidu's AI hasn't yet grown large enough to change the direction of the entire company.
Kuaishou is a completely different story.
Kling is perhaps the most unexpected card in this round of AI competition among Chinese internet giants.
In Q2, Kling AI revenue had already exceeded 850 million yuan, with year-on-year growth exceeding 200%.
This is a very impressive number.
It means Kling has crossed the hardest hurdle for many AI products: moving from having a strong model, many users, and a vibrant community to actually "having people willing to pay."
But after the earnings report came out, Kuaishou's stock still plummeted.
Because looking at the entire group, Kuaishou's revenue only grew 1.4%, and adjusted net profit also declined significantly.
Thus, a very interesting picture emerges:
A star AI product growing over 200% cannot yet save a group growing only in the single digits.
A new business being "very good" and it being "large enough to change the company" are two completely different things.
Xiaomi is taking a fifth path.
At least for now, Xiaomi's true second growth curve is still not AI, but automobiles.
In Q2, revenue from smart electric vehicles and AI and other innovation businesses reached 24.9 billion yuan, of which automobile revenue was 23.9 billion yuan.
So comparing "Xiaomi's AI revenue" directly with Alibaba Cloud or Kling doesn't make much sense in itself.
The significance of AI for Xiaomi is more like an underlying capability gradually permeating phones, cars, home appliances, and robots.
What Xiaomi is really betting on is another question:
If phones, cars, home appliances, and robots all become AI terminals in the future, will the one with the most hardware gateways conversely possess a huge AI advantage?
This is also why MiMo is most important for Xiaomi.
It doesn't necessarily have to become an independent large model business.
If AI can ultimately truly connect phones, cars, IoT, and robots into one system, then the "Human x Car x Home Full Ecosystem" would for the first time not just be about connecting devices, but could potentially share a common brain.
Of course, this path is also very expensive.
With several battlefronts—the smartphone cycle, automobile manufacturing, chip R&D, and AI investment—simultaneously underway, Xiaomi's story is also becoming increasingly heavy.
It is the least traditional internet company among the six, and may therefore have the most different AI outcome.
Finally, there's NetEase.
NetEase might be the most unique one among the six companies.
It is certainly also using AI.
AI can generate art, assist in program development, improve NPC interaction capabilities, and reduce content production costs.
But at least for now, NetEase isn't in a particular hurry to repackage itself as an "AI company."
In Q2, revenue from games and related value-added services was 25.0 billion yuan, a year-on-year increase of 9.7%. Products like Where Winds Meet and Fantasy Westward Journey still drive the company's core growth.
This actually reminds us of a question easily overlooked in the AI hype:
Not every internet company's value needs to be reinterpreted by AI.
03
For a game company, AI is certainly important.
But what ultimately determines whether a game succeeds is still gameplay, content, aesthetics, and operations.
If AI ultimately just makes NetEase 30% more efficient at making games, shortens production cycles by 20%, and makes NPCs a bit smarter, it might already be creating immense economic value.
To some extent, AI *not* being the most prominent line in NetEase's financial report might itself be an answer.
So, placing these six companies side by side, you'll find that the so-called "Chinese internet entering the AI era" is actually an overly general statement.
They are not entering the same AI era at all.
Tencent faces the problem of return on capital.
Alibaba faces the problem of reinvestment.
Baidu faces the problem of replacing old businesses with new ones.
Kuaishou faces the problem of scale.
Xiaomi faces the problem of the ecosystem.
NetEase even faces another question: Do I even need to redefine myself because of AI?
Six companies, six completely different sets of questions.
What will truly determine these companies' future value is what exactly happens after AI lands within their original business models.
Tencent's originally strong cash flow is further amplified by AI, but also consumed by it;
Alibaba's original cloud computing infrastructure regains strategic value because of AI, while simultaneously requiring massive investment;
Baidu's original dependence on search is exposed even more clearly in the AI era;
Kuaishou, originally lacking a second growth curve, unexpectedly got a new card with Kling;
Xiaomi's originally vast hardware ecosystem could for the first time be truly connected because of AI;
NetEase's originally core content capabilities remind everyone that sometimes AI doesn't need to be the protagonist.
So, AI has not pulled the Chinese internet back to the same starting line.
It goes back to the initial judgment: The stronger the AI consensus, the harder it is to capture a premium for AI itself.
When models, computing power, and Agents gradually become standard configurations for major players, the market will eventually return to that most ancient question:
Is the original business good?
Can you turn AI into revenue, profit, or higher efficiency?
What price must be paid to achieve this?
In the end, AI may not be a card that reshuffles the deck.
It is more like a magnifying glass.
It magnifies each company's cash flow, old businesses, organizational capabilities, historical baggage, and true competitive advantages.
Thus, in the same AI era, there are six different fates.
This article is from WeChat public account "Wu Duidui" (ID: esnql520), author: Wu Duidui
