[Summary]

Chinese AI companies appear to be technologically strong.

So, is it really profitable as a business?

Here are the questions investors should be asking in 2026. It is not about the performance of the AI ​​model or the number of tokens used, but how to turn AI into profit margin and free cash flow (FCF).

After DeepSeek, price competition for model APIs has become more likely to occur in the Chinese AI market. Simply selling AI externally may reduce profit margins even if sales increase.

On the other hand, companies that can incorporate AI into their advertising, payments, e-commerce, business SaaS, and industrial infrastructure can use AI to improve the profitability of existing businesses rather than generating additional sales.

To conclude first, Tencent is strong in terms of current FCF and profit margin stability. Alibaba has a lot of room for growth in the AI ​​cloud. Baidu is waiting to pivot from search advertising to AI business. Huawei is unlisted and a detailed comparison is difficult, but its AI infrastructure for governments and state-owned enterprises puts pressure on the margins of the three listed companies.

However, the winner has not yet been determined.

What we should be looking at in the second half of 2026 is not AI sales, but the quality of profit margins and the speed of FCF return.

Companies that sell AI and companies that use AI are different

When comparing Chinese AI companies, the most important thing is to distinguish between "Are they selling AI" and "Are they using AI?"

Companies that sell AI provide clouds, APIs, models, AI chips, and enterprise solutions to external customers. Sales are likely to grow, but are susceptible to price competition, capital investment, GPU/NPU utilization rates, and power costs.

Companies using AI will incorporate it into their advertising, e-commerce, payments, games, business tools, logistics, and industrial systems. Although it is difficult to see the sales of AI alone, it is effective in improving the gross profit margin and conversion of existing businesses.

This difference results in the difference between profit margin and FCF.

CompaniesSelling AIUsing AIPoints to check in the second half of 2026
Tencent△◎Ad unit price, Business Services, and FCF sustainability
Alibaba◎○Cloud EBITA margin, CAPEX, FCF recovery
Baidu◎△Search advertising bottoming out, AI Cloud profitability, Apollo Go
Huawei◎○Although detailed comparisons are difficult as the company is unlisted, recruitment trends in government/state-owned enterprise projects

The reason why Tencent looks strong here is because its structure is to mix AI into existing businesses rather than selling AI itself.

Alibaba and Baidu are more likely to sell AI cloud services externally. Sales growth is impressive, but the investment burden also comes to the fore.

Huawei is a little special. Since it is unlisted, it is difficult to directly compare profit margins and AI standalone FCF, but by combining Ascend, Huawei Cloud, and Pangu, we will go deeper into government and state-owned enterprises. This will affect the profit distribution of the entire Chinese AI market.

Comparison of 4 companies: Current position of profit margin and FCF

When looking at the latest public data from an investor's perspective, the views are quite different.

CompanyHow to see profit margin/FCFStrengthsChallenges
Tencent◎Non-IFRS operating profit margin 38.5%, FCF 56.7 billion yuan. Existing businesses have strong cash-generating abilityAI cloud purity is low, and it is difficult to see a growth story for AI alone
Alibaba△Cloud sales increased by 38%, AI-related product sales grew by 3 digitsCompany-wide FCF leaked. CAPEX and user acquisition costs are heavy
Baidu△ to ×AI-powered Business increased by 49%, AI ratio increasedSearch advertising decrease and FCF negative. Waiting for reverse numbers
Huawei○Company-wide operating cash flow is high. Strong in AI infrastructure and industrial projectsUnlisted, AI standalone profit margin and FCF are difficult to compare

If you just look at this table, Tencent is the strongest.

However, this is a matter of "current financial stability" and does not determine future stock price momentum.

Once Alibaba overcomes the investment trough, AI Cloud's operating leverage may be effective.

Baidu will depend on the extent to which AI cloud, AI advertising, and Apollo Go can replace the decline in search advertising.

Although Huawei is not a direct investment target, it will change the competitive conditions for AI infrastructure for governments and state-owned enterprises.

Tencent: Increase the profit margin of existing businesses rather than selling AI

Tencent has the most obvious cash flow as of 2026.

2026 1Q sales were 196.458 billion yuan, non-IFRS operating profit was 75.6 billion yuan, and non-IFRS operating profit margin was 38.5%. FCF was 56.7 billion yuan, an increase of 20% year-on-year.

Tencent's strength lies in the fact that it does not have to forcefully sell AI as an externally sold cloud.

There are existing high-frequency contact points: WeChat, advertising, games, payments, mini-shops, and Business Services. AI rides on it.

In advertising, if the AI ​​recommendation model is effective, it will be reflected in the ad unit price and conversion. In mini-shops, it is effective for store management and product recommendation. Business Services will lead to cloud and enterprise services, including AI-related demand.

In other words, it is easier to understand Tencent's AI as ``improving the profit margin of existing businesses'' rather than as a ``sales item.''

Numbers to look at in TencentReasons
Marketing Services Growth RateIs AI Advertising Boosting Unit Prices and Efficiency
Business Services Growth RateIs AI Cloud/AI Agent demand monetizing?
Non-IFRS operating profit marginCan a high level be maintained even after absorbing AI investment?
FCFWill there be cash left after investing in AI?

The weak point is that it is easy to see that the purity of the company as an AI cloud company is low.

Alibaba may be easier to understand for investors aiming for a major revaluation through external sales of AI infrastructure.

Still, Tencent is the most secure of the four companies in terms of short- to medium-term financial stability.

Alibaba: AI cloud grows, but FCF is in investment trough

Alibaba is one of the favorites for China's AI cloud.

Cloud Intelligence Group's revenue in Q4 of the fiscal year ending March 2026 was 41.626 billion yuan, an increase of 38% compared to the same period last year. AI-related product sales were 8.971 billion yuan, representing 11 consecutive quarters of triple-digit growth year-on-year.

In terms of sales growth alone, it's pretty strong.

The problem is cash.

Group adjusted EBITA for the quarter was 5.102 billion yuan, down 84% year-on-year. FCF had an outflow of 17.3 billion yuan. The company attributes the decline in FCF to Quick Commerce, Qwen app user acquisition, and cloud infrastructure spending.

The issue for Alibaba is not whether AI is growing.

The question is when will AI cloud turn into a high-margin business?

Alibaba reversal conditionsReasons to watch
Cloud EBITA margin improvementAre AI sales profitable?
CAPEX peaks outWill capital investment continue to put pressure on FCF
Conversion from Qwen to DingTalk/SaaSIs it possible to move from API sale to continuous billing
Recovery of company-wide FCFHas AI investment started to return as cash

Alibaba has operating leverage.

If you build the infrastructure first, get customers onboard, and increase usage rates, the weight of fixed costs may turn into an improvement in profit margins.

However, that is still being confirmed.

In the second half of 2026, I would like to see whether the increased sales return to EBITA and FCF rather than Cloud's growth itself.

Baidu: AI transformation is progressing, but there are serious holes in search advertising

Baidu's structural transformation is the easiest to understand among the four companies.

It's moving from a search advertising company to an AI company, including AI Cloud, AI Advertising, and Apollo Go.

Baidu Core AI-powered Business in 1Q 2026 was 13.6 billion yuan, an increase of 49% compared to the same period last year. AI Cloud Infrastructure was 8.8 billion yuan, an increase of 79% year on year. Meanwhile, online marketing sales were 12.6 billion yuan, down 22% year-on-year.

This is the difficulty with Baidu.

AI is growing.

However, search advertising, which had a high profit margin, is weak.

Even if AI cloud, which tends to have low margins, grows, it won't necessarily replace advertising profits right away.

Operating cash flow was a positive 2.670 billion yuan, but FCF after deducting capital expenditures was a negative 3.246 billion yuan on a Baidu Inc. basis.

Looking at Baidu, changes in sales composition alone are not enough.

Items to check on BaiduReasons to watch
Decrease rate of search advertisingHow much damage is being done to existing high-profit businesses
AI Cloud profitabilityIs growth accompanied by profits
AI-native Marketing ServicesWill AI advertising fill the gap in search advertising
Apollo GoDoes the unit economics improve based on the number of rides?
FCFHas AI conversion started to generate cash

Baidu's appeal is that it has room for reversal.

But reversal requires numbers.

What the market wants to see is not an increase in AI sales, but whether the decline in search advertising can really be replaced by profits and FCF from the AI ​​business.

Huawei: unlisted but national infrastructure company changing profit distribution

Huawei is not a listed company.

Therefore, it is not possible to compare quarterly AI profit margins and FCF in the same way as Tencent, Alibaba, and Baidu.

Still, if Huawei is not included in the overview of China's AI market, an important part will be missed.

Huawei's essence is not so much an AI company as a national infrastructure company in the AI ​​era.

We bring together Ascend chips, Huawei Cloud, Pangu, CANN/MindSpore, and communication infrastructure for governments, state-owned enterprises, and critical infrastructure.

In 2025, Huawei had sales of 880.941 billion yuan, operating profit of 96.937 billion yuan, and operating cash flow of 127.384 billion yuan. On a company-wide basis, the company has sufficient cash generation ability.

However, the profit margin and FCF of AI alone cannot be seen in detail. This is a major constraint when compared to the three listed companies.

For investors, Huawei is not something to buy directly, but something that changes the margins of other companies.

Areas of pressure from HuaweiImpact
Government/state-owned enterprise cloudAlibaba, Tencent, and Baidu are finding it difficult to take on more projects
Domestic AI chipsReduce dependence on NVIDIA and become a standard candidate for domestically produced AI infrastructure
Industrial AgentDeep-dive into B2B/B2G projects with Pangu and Huawei Cloud
Developer baseAs CANN/MindSpore spreads, the lock-in of cloud usage will become stronger

The stronger Huawei becomes, the more the profit distribution in China's AI market will change.

Particularly in B2G, public infrastructure, and cloud for state-owned enterprises, this could reduce the growth potential and profit margins of the three private tech companies.

Roadmap that investors should see

In the 2026 snapshot, Tencent's financials are the most stable.

However, if we fix our conclusion only on that point, we will overlook Alibaba's operating leverage and Baidu's potential for reversal.

Each company's roadmap can be organized as follows.

CompaniesWhat to see in the second half of 2026
TencentCan you maintain profit margins and FCF in advertising, payments, and Business Services even if you increase AI investment?
AlibabaWill FCF start to return due to Cloud EBITA margin improvement and CAPEX peaking out
BaiduCan we see the bottoming out of search advertising and the profit contribution of AI Cloud, AI advertising, and Apollo Go?
HuaweiTo what extent will the adoption of AI infrastructure for governments and state-owned enterprises put pressure on the market share and profit margins of private cloud companies?

In the end, what we should be looking at is not AI sales.

It is the quality of profit margin.

Even for the same AI sales, the value is different between sales that grew due to API price competition and sales that grew due to SaaS billing and advertising efficiency improvements.

Furthermore, the bottom line is FCF rather than profits.

AI investments tend to be initially viewed as CAPEX, human resources, data center, and inference costs. If that investment is to produce value, it will eventually have to come back to operating CF and FCF.

Conclusion: Winners for AI companies are determined by cash, not sales.

Winners among Chinese AI companies will not be determined solely by model performance or AI sales.

Will companies selling AI be able to overcome price competition and capital investment and move to recurring billing SaaS and high-value-added cloud services?

How much can companies using AI lower the CAC of existing businesses and increase the profit margins of advertising, payments, EC, and business services?

At the moment, Tencent has the most stable profit margin and FCF.

However, once Alibaba overcomes the investment trough, operating leverage may come into play. Baidu has room to reevaluate if the results of its structural transformation are reflected in FCF. Although unlisted, Huawei continues to change the distribution of profits in China's AI market as a national infrastructure company.

Investors shouldn't be looking for flashy AI announcements every quarter.

Profit margin, CAPEX, operating CF, FCF.

And the speed at which AI turns into cash rather than sales.

This article is intended to summarize the thinking behind investment decisions, and is not intended to recommend buying or selling specific stocks. Huawei is a private company, and general investors cannot directly buy or sell its shares. Chinese stocks, Hong Kong stocks, and US-listed ADRs are subject to risks associated with price fluctuation risk, exchange rate risk, liquidity risk, regulatory risk, geopolitical risk, and differences in accounting and disclosure systems.

Related articles

Source

This article is for educational and informational purposes only, based on public information. It is not a recommendation or solicitation to buy or sell any specific security or financial product. Although care is taken with accuracy, the content and future investment outcomes are not guaranteed. Final investment decisions should be made at your own judgment and responsibility.