[Summary]
When looking at the Chinese AI market, there are things that are overlooked if you only follow Alibaba, Tencent, and Baidu.
That's Huawei.
Huawei is a private company, so ordinary investors cannot directly buy its shares. Still, it cannot be ignored when considering Chinese AI investment. The reason is that it is in a position to vertically connect the Ascend chip, Huawei Cloud, Pangu, MindSpore/CANN, and communication infrastructure to meet the AI demand for governments, state-owned enterprises, and critical infrastructure.
Huawei's essence is not so much an AI company as a national infrastructure company in the AI era.
For investors, Huawei is more than just an AI company.
It is a competitive factor that will change the profit distribution in China's AI cloud market and could put pressure on the margins of Alibaba Cloud, Tencent Cloud, and Baidu AI Cloud.
In this article, we will summarize Huawei's Ascend, Pangu, national strategy, monetization route, and risks that investors should consider as of June 1, 2026.
Huawei is not listed. Why investors still watch
I would like to check this first.
Huawei is not a listed company.
Therefore, unlike Alibaba, Tencent, and Baidu, its valuation cannot be directly compared on the stock market.
However, just because a company is unlisted does not mean it is not important.
In China's AI market, profit margins are determined not only by who wins with models, but also by who controls computing resources, cloud computing, industrial customers, and government projects.
Among these, Huawei is strong in computing resources and industrial customers.
In 2025, Huawei had sales of 880.941 billion yuan, net profit of 68.036 billion yuan, and operating cash flow of 127.384 billion yuan. R&D investment was 192.3 billion yuan, accounting for 21.8% of sales.
This R&D capacity and customer base in telecommunications, cloud, terminals, and industrial infrastructure make Huawei a ``huge pressure that is hard to see from the outside'' in China's AI market.
Ascend: Receptacle for NVIDIA alternative demand
The starting point of Huawei's AI strategy is the Ascend chip.
Due to US export restrictions on high-performance GPUs, it is becoming difficult for Chinese companies to freely procure NVIDIA's top-of-the-line GPUs. Therefore, the importance of domestically produced AI chips has increased.
Ascend is one alternative candidate to fill this void.
However, don't get too excited here.
This is not to say that Ascend completely replaced NVIDIA. NVIDIA still has a strong wall in terms of CUDA's developer ecosystem, software maturity, performance stability, and global adoption record.
Still, there is strong demand within China for ``usable domestic AI infrastructure.'' Governments, state-owned enterprises, telecommunications, finance, and public infrastructure emphasize not only technical performance but also supply stability, data management, security, and procurement policies.
Huawei comes in here.
| Items viewed by investors | Meaning |
|---|---|
| Ascend chip | Core candidate for domestic AI computing resources |
| Atlas/SuperPoD | Infrastructure for large-scale AI learning and inference |
| CANN/MindSpore | Software platform against CUDA |
| Huawei Cloud | Collection device that connects chip demand to cloud usage |
What's scary about Huawei is that it doesn't just sell chips.
Using Ascend as an entry point, servers, clouds, development environments, and industrial AI solutions can be sold all at once.
This is where we differ from individual semiconductor manufacturers.
MindSpore and CANN: Can we create a developer ecosystem?
NVIDIA's strength is not only in GPU performance.
It has a strong developer ecosystem centered around CUDA.
Huawei also understands this. We are developing CANN, MindSpore, ModelArts, AI Gallery, etc. for Ascend to enable developers to create AI in the Huawei environment.
Huawei Cloud's Ascend AI Cloud Service focuses on toolchains for Ascend, migration of open source models, and integration of AI agents.
This is not just a technical explanation.
For investors, it's a matter of lock-in.
Developers build in the Huawei environment
↓
Models and agents are optimized for Ascend
↓
Huawei Cloud and ModelArts usage increases
↓
This leads to cloud fees, maintenance, and additional development
The more this loop continues, the closer Huawei's AI infrastructure will be to stock-type profits.
On the other hand, there are also challenges.
In terms of the natural spread of private startups and global developers, there are times when Alibaba's Qwen and international open source AI infrastructure appear to be stronger. Huawei is making CANN open and supporting major frameworks, but it will take time to break global standards like CUDA.
Therefore, when looking at MindSpore/CANN, I want to see not only the strength of the presentation materials, but also how much developers actually use it and how much it leads to paid cloud usage.
Pangu: Aiming for industrial agents rather than consumer AI
Huawei's Pangu is easier to understand as an industrial AI than as a ChatGPT-type AI for general consumers.
For Pangu Models 5.5, Huawei Cloud emphasizes industrial applications such as agriculture, manufacturing, and scientific research. Furthermore, by combining ModelArts and AI Cloud Service, it provides a platform for companies to create specialized models using their own data.
What we are aiming for here is not just chat usage fees.
The idea is to incorporate AI into business systems in areas such as factories, mines, electricity, finance, communications, and urban infrastructure.
| Pangu's aim | Form of monetization |
|---|---|
| Manufacturing/Electricity | Anomaly detection, maintenance, demand forecasting, equipment optimization |
| Finance | Risk management, document processing, customer relations, internal audit |
| Communications | Network operation, failure response, autonomous operation |
| Public/Urban | Urban management, transportation, administrative systems, data analysis |
This area is not fancy.
However, the unit price tends to be high. Switching costs after introduction are also high. If maintenance, cloud usage, and additional development continue, we will approach recurring revenue.
This is why Huawei is strong in the AI agent market.
Rather than explosively popularizing it for consumers, it penetrates deep into the industry.
National budget and domestic substitution: Huawei's biggest tailwind
Huawei's biggest tailwinds are national security and domestic alternatives.
China continues to move toward reducing its dependence on U.S. technology in the digitalization of critical infrastructure, communications, finance, government, and state-owned enterprises.
This trend is strong for Huawei.
U.S. export controls and geopolitical risk
↓
Need for domestic AI infrastructure
↓
Procurement demand from government, SOEs, and critical infrastructure
↓
Ascend + Huawei Cloud + Pangu + operations and maintenance
↓
Long-term projects, additional development, and cloud usage
What I would like to caution here is not to declare that Japan has a monopoly on the national budget.
In reality, multiple domestic vendors are involved in government-related projects, and procurement sources are divided by local government and industry.
However, it is certain that Huawei is in an advantageous position. We have a proven track record in communications infrastructure, cloud, chips, AI models, terminals, security, and a sales network for industrial customers, making it easy to make full-stack proposals.
Investors should be looking at more than just how much of national demand Huawei captures.
The more Huawei captures, the more space will be created that will be difficult for Alibaba Cloud, Tencent Cloud, and Baidu AI Cloud to capture.
Where does Huawei make money?
Huawei's AI-related revenue can be easily understood by looking at it roughly in three layers.
| Layers | Main products and services | Nature of earnings | Investor perspective |
|---|---|---|---|
| Hardware | Ascend chips, Atlas, servers, communication equipment | Flow type, replacement type | High unit price, but subject to parts, manufacturing, and supply constraints |
| Platform | Huawei Cloud, ModelArts, CANN, MindSpore | Pay-as-you-go, subscription, development platform | As usage increases, stock becomes available |
| Applications | Pangu, industrial agent, operation and maintenance, industry solutions | Initial implementation + maintenance + additional development | LTV tends to be longer with deep industrial development |
When these three layers are connected, Huawei is strong.
sell chips
Run in the cloud.
Put Pangu and Agent into work.
Continue maintenance and additional development.
This trend is a much more investor-friendly profit structure than a one-off AI boom.
Huawei's strength lies not only in its sales volume, but also in the fact that it can easily maintain relatively high profit margins through high-priced government and state-owned enterprise projects. If profit margins are maintained, it will be easier to see how continued use of Huawei Cloud and Pangu will lead to FCF.
However, Huawei is unlisted, and its AI-related profit margins and FCF are not disclosed in detail. Therefore, it is dangerous to conclude that ``AI generates huge FCF.''
What you should look at is the company's overall operating cash flow, R&D investment, cloud sales, ICT infrastructure sales, and the adoption record of domestically produced AI infrastructure.
Impact on Alibaba, Tencent, and Baidu
If Huawei becomes stronger, what will happen to listed Chinese AI companies?
The most obvious impact is on cloud profit margins.
| Companies | Pressure from Huawei |
|---|---|
| Alibaba | Competitive pressure on AI cloud projects for governments and state-owned enterprises. Differentiation is needed in the private sector/developer ecosystem |
| Tencent | WeChat economic zone is strong, but Huawei is likely to gain an advantage in public and industrial cloud |
| Baidu | Direct competition between AI cloud and domestic GPU cloud. Differentiation other than Apollo Go is in question |
Alibaba is strong with Qwen and private developers.
Tencent is strong in WeChat and existing FCF.
Baidu differentiates itself with AI transformation and Apollo Go.
However, Huawei has a strong presence in the areas of government, state-owned enterprises, and critical infrastructure.
For this reason, when investors look at China's top three AI companies, it is not enough to simply look at the growth rate of AI sales.
In which customer segments is it in direct conflict with Huawei?
In what areas can you avoid Huawei and still make profits?
If you don't look at it that far, you will misread the AI cloud margin.
Risk: Huawei also has weaknesses
Huawei is strong.
However, there are also weaknesses.
Unlisted with limited disclosure
Because Huawei is unlisted, it is difficult to closely track profit margins by division and FCF for AI cloud alone, unlike publicly traded cloud companies.
Investors can only read indirectly from company-wide numbers and official announcements, orders from surrounding companies, cloud market share, and semiconductor supply chain.
Developer Ecosystem Walls
Even if Ascend and CANN grow, it will not be easy to replace global standards like CUDA in the short term.
In an environment where private startups can experiment freely and expand at low cost, Alibaba, ByteDance, and open source platforms are sometimes chosen.
Geopolitics and Supply Chain
While Huawei enjoys a tailwind from domestic substitution, it is also at the center of geopolitical risks.
Multiple constraints remain for AI chips, including high-performance memory, advanced manufacturing, EDA, equipment, and packaging. As Ascend grows, the strength of its supply network will be tested.
Conclusion: Huawei is changing the profit margins of China's three AI giants
Huawei is not a "fourth listing candidate" in China's AI market.
Huawei's essence is not so much an AI company as a national infrastructure company in the AI era. Through AI chips, the cloud, and industrial agents, it is trying to capture China's demand for critical infrastructure itself.
Although it remains unlisted, it is changing the profit margins of listed AI companies.
Conserve computational resources with Ascend.
Engage developers with CANN/MindSpore.
Charge usage with Huawei Cloud.
Deep dive into governments, state-owned enterprises, and critical infrastructure with Pangu and Industrial Agent.
As this vertical integration progresses, the distribution of profits in China's AI market will change.
When investors evaluate Alibaba, Tencent, and Baidu, it's dangerous not to look at Huawei.
In particular, if you look at the profit margin of AI cloud, projects for governments and state-owned enterprises, domestically produced GPU cloud, and monetization of industrial agents, Huawei will definitely be included in the comparison.
Huawei itself cannot be purchased directly.
Still, where will Huawei absorb profits and where will it cut into competitors' margins?
If you look at that, the way you look at Chinese AI stocks will change considerably.
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.
FAQ
Is Huawei listed?
Huawei is a private company, and general investors cannot buy or sell its shares directly.
However, Huawei's move could affect the competitive environment for listed Chinese AI companies such as Alibaba, Tencent, and Baidu in terms of cloud revenue, semiconductor procurement, and projects for government and state-owned enterprises.
What is Ascend?
Ascend is Huawei's brand of AI chips and AI computing infrastructure.
In China, it is attracting attention as an alternative candidate for AI learning and inference due to export restrictions on high-performance GPUs and the trend toward domestically produced alternatives.
Is Huawei Cloud strong?
Huawei Cloud has a presence in areas where security and domestic substitution are important, such as government, state-owned enterprises, telecommunications, finance, and public infrastructure.
On the other hand, competition from Alibaba Cloud and open source infrastructure remains in private startups and the global developer community.
What is Pangu?
Pangu is a group of large-scale industrial AI models deployed by Huawei Cloud.
Rather than chat AI for general consumers, it is more likely to be used to improve business operations in manufacturing, finance, electricity, communications, urban infrastructure, etc., or to be applied as an industrial agent.
Related articles
- [Who is the winner in China's AI agent market? Compare the monetization strategies of Tencent, Alibaba, and Huawei] (/strategy/2026/06/01/china-ai-agent-market-tencent-alibaba-huawei-2026.html)
- Comparing China AI cloud market | Who is the winner among Alibaba, Huawei, Tencent, and Baidu?
- Comparing China's Big Three AI | Who is the winner among Alibaba, Tencent, and Baidu?
- [Is it time to buy Alibaba stock? Analyzing the future from AI cloud, Qwen, and FCF [2026 edition]](/strategy/2026/05/31/alibaba-ai-cloud-qwen-fcf-stock-outlook-2026.html)
Source
- Huawei「2025 Annual Report」 https://www.huawei.com/en/annual-report/2025
- Huawei「Huawei Releases 2025 Annual Report: Performance in Line with Forecast」 https://www.huawei.com/en/news/2026/3/annual-report-2025
- Huawei「Huawei Unveiled the Latest SuperPoD, Making an AI Infrastructure New Option to the World」 https://www.huawei.com/en/news/2026/3/mwc-superpod-ai
- Huawei Cloud「Huawei Cloud Announces Pangu Models 5.5 and All-new AI Cloud Service」 https://www.huaweicloud.com/eu/news/20250620192415143.html
- Huawei Cloud「Ascend AI Cloud Service」 https://www.huaweicloud.com/intl/en-us/product/modelarts/ascendaicloud.html