AnswersPublished15 sources
How Huawei, Alibaba, ByteDance and Tencent Are Approaching the AI Phone
As of October 10, 2026, the clearest contrast is Huawei’s push across chips, software and on device models versus more partnership oriented approaches attributed to Alibaba, ByteDance and Tencent. The key business challenge is getting phone makers, AI providers and major apps to share access and value when an assist...
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- As of October 10, 2026, the clearest contrast is Huawei’s push across chips, software and on device models versus more partnership oriented approaches attributed to Alibaba, ByteDance and Tencent.
- The key business challenge is getting phone makers, AI providers and major apps to share access and value when an assistant completes tasks across services.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: As of October 10, 2026, how are Huawei, Alibaba, ByteDance and Tencent pursuing different strategies to control the AI layer between Chinese. Article summary: The strategic contrast is full-stack control for Huawei, a focused AI-phone partnership for ByteDance, a cross-brand intelligence platform for Alibaba, and service-level integration for Tencent. However, the available ex. Topic tags: general, news, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts w
The contest to put an AI assistant between smartphone users and their apps is also a contest over control: who supplies the intelligence, who owns the user relationship, and which services an assistant can reach. An October 9 analysis frames Huawei, Alibaba, ByteDance and Tencent as taking different routes—but the available reporting verifies Huawei’s hardware direction more clearly than the detailed partnerships attributed to the other companies. 13
Four different routes to the AI entry point
Huawei: build across the stack
Huawei’s approach is the most vertically integrated of the four as described in the available material. The company’s Mate 90 launch materials describe HarmonyOS 7 as part of an integrated hardware, software, chip and cloud stack. Huawei also said the Mate 90 series can run a 30-billion-parameter mixture-of-experts model on-device. 3
12
That is a company-reported capability, not an independent benchmark of sustained performance, power use or real-world task quality. Nor does on-device inference mean every assistant function works offline: tasks that depend on online services still require connectivity.
The chip’s supply chain is another constraint. Reuters reports that premium Mate 90 models use the Kirin 9050 Pro and that Huawei did not disclose its manufacturer. TrendForce’s report flags the design’s greater wafer requirements. 1
2 Those details make supply and production relevant to Huawei’s strategy, but they do not establish the chip’s long-term availability or manufacturing yield.
Alibaba: extend intelligence across devices
The October 9 analysis characterizes Alibaba’s strategy as broader and platform-oriented. 13 In principle, supplying AI capabilities to multiple phone makers could extend a model platform across brands without requiring the provider to build a phone business of its own.
That approach also leaves important questions about control: which company handles system permissions, presents the assistant to users, and keeps the customer relationship? The sources available here do not independently confirm the specific Qwen deployments or division of responsibilities described in the research prompt, so those details should not be treated as established facts.
ByteDance: pursue a focused handset partnership
The same analysis labels ByteDance’s route as focused. 13 A close partnership with a handset maker could give an assistant a more prominent role on selected devices, but broader app access would still matter: an assistant that can act only within a limited set of services cannot reliably complete users’ tasks across the phone.
The supplied sources do not verify the specific Nubia handset, a shift from screen automation to MCP or A2A calls, or the claimed limits on which services its automation can use. These remain unconfirmed details rather than a basis for judging the partnership’s actual capabilities.
Tencent: work through services rather than own the system assistant
The analysis also describes Tencent’s approach as service-level rather than a direct contest for the phone’s primary assistant. 13 This distinction matters because an AI agent can be the system-wide entry point while relying on a separate service—such as a messaging or commerce app—to fulfill a request.
However, the available sources do not verify the named A2A partnerships or the reported Qualcomm deployments of Hunyuan models. What can be said more generally is that access to services is a separate question from which company owns the initial interaction.
Why the business model may decide who gets access
An AI assistant can change how users discover and use apps. That creates potential friction among three groups:
- Phone makers may worry that an outside assistant will own the interaction and leave them mainly selling hardware.
- Model providers can earn from API usage without necessarily owning the customer relationship or receiving a share of the value from a completed transaction.
- Apps and super apps may be essential to fulfilling requests, while also competing with the assistant for user attention and control of the customer journey.
The October 9 analysis identifies app access and revenue sharing as unresolved parts of the contest. 13 The available reporting does not establish that major apps blocked a particular handset, so that claim cannot be confirmed here.
Why payment for completed tasks is worth considering
Sharing revenue when an assistant completes a verified task—such as a booking or purchase—could align incentives more directly than either advertising exposure or token consumption. Advertising measures attention; token counts measure computation. Neither necessarily shows that the user got what they wanted.
A completed-task model could give the assistant provider, phone maker, model supplier and service a reason to cooperate. But it would require clear rules for user permission, attribution, cancellations and disputes. It is a plausible approach, not a proven requirement or an established industry agreement.
What is known—and what remains open
The broad strategic framing is clear: Huawei is pushing into its own device stack, while the other companies are described as favoring different forms of platform, handset or service integration. 13 Huawei’s Mate 90 chip and on-device-model claims have supporting launch coverage, though performance and supply questions remain.
1
2
3
12
For the specific ByteDance, Alibaba and Tencent deployments, the supplied sources do not provide enough verification to state the details as fact. And until phone makers, AI providers and apps agree on access and how value is shared, no single approach can be assumed to have secured the phone’s primary AI entry point.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "How Huawei, Alibaba, ByteDance and Tencent Are Approaching the AI Phone"?
As of October 10, 2026, the clearest contrast is Huawei’s push across chips, software and on device models versus more partnership oriented approaches attributed to Alibaba, ByteDance and Tencent.
What are the key points to validate first?
As of October 10, 2026, the clearest contrast is Huawei’s push across chips, software and on device models versus more partnership oriented approaches attributed to Alibaba, ByteDance and Tencent. The key business challenge is getting phone makers, AI providers and major apps to share access and value when an assistant completes tasks across services.