US chip restrictions may delay Tencent’s next model


Chinese tech giant Tencent's leadership stated that new US restrictions on remote access to advanced processors will likely slow the development of its next artificial intelligence model, without substantially impacting its current cloud revenue. This assessment followed the announcement of third-quarter revenue reaching 207.3 billion yuan and a surge in capital expenditure to 81.4 billion yuan.
Tencent leadership’s remarks draw a clear technical and commercial distinction between training frontier models and running cloud inference services, illustrating how restrictions on remote access to advanced US processors could slow future research without disrupting current cloud revenue.
Tencent leadership stated that new US restrictions on remote access to advanced processors will likely slow the development of its next AI model, while leaving the company’s current cloud computing revenue relatively unaffected. This assessment was made during a discussion following the earnings announcement on November 18, as reported by Reuters on November 18, 2026, at 08:43 UTC (12:43 Gulf Standard Time).
This operational approach is based on a precise distinction between two different technical activities in the AI lifecycle: training new frontier models on the one hand, and fulfilling user and customer requests using pre-trained models on the other.
Tencent Chief Strategy Officer James Mitchell explained that a significant portion of the company’s cloud activity centers around inference operations, tasks that can be efficiently run on fewer domestic Chinese processors when software models are effectively optimized. Meanwhile, company President Martin Lau noted that the loss of access to more powerful computing capabilities could constrain future system training operations in the near term, while encouraging the development of China’s chip supply chains over time. These insights remain within the framework of executive management expectations rather than proven technical results on the ground.
The company’s financial results showed third-quarter revenue reaching 207.3 billion yuan (approximately $31.1 billion), an 8% increase compared to the previous year, though missing London Stock Exchange Group expectations of 210.8 billion yuan. Revenue from fintech and business services, the segment that includes cloud operations, grew 9% to 63.7 billion yuan, while net income rose 0.8% to 63.61 billion yuan, beating expectations. These financial indicators cover multiple business activities, so the performance of the entire segment cannot be attributed to AI alone.
The company’s capital expenditure jumped to 81.4 billion yuan in the September quarter, up from 52.8 billion yuan in the preceding June quarter. However, the scale of capital expenditure does not in itself prove the availability of additional computing capacity that is fully installed and utilized for AI applications, as Tencent still needs to purchase, integrate, and actually operate systems to meet customer demand.
The remote computing restrictions announced earlier in the week pertain to accessing advanced US hardware via overseas facilities and installations. Tencent’s statements provide new company-specific evidence regarding the potential impacts of these rules; inference services can continue operating using currently available hardware even as advanced training becomes more difficult, which explains management’s expectation of a limited immediate revenue impact while warning of a slowdown in research progress. The report highlights a less visible impact of semiconductor restrictions on computing services as well as direct chip shipments, and upcoming evidence will include Tencent’s model release timeline, cloud customer revenue, actual domestic chip utilization, and the financial return on infrastructure expansion.
Q: Why does Tencent expect its next model development to be affected without its current cloud revenue being harmed?
A: Because most of its cloud activity relies on inference to run pre-trained models, which can be executed on fewer domestic processors with optimized efficiency, whereas training new models requires superior computing power restricted by US rules.
Q: What are Tencent’s standout financial figures for the third quarter of 2026?
A: Revenue reached 207.3 billion yuan, up 8% but below estimates of 210.8 billion yuan. Fintech and business services grew 9% to 63.7 billion yuan, while net income hit 63.61 billion yuan.
Q: What level did Tencent’s capital expenditure reach in the September quarter?
A: Capital expenditure rose to 81.4 billion yuan compared to about 52.8 billion yuan in the June quarter, noting that financial spending alone does not mean computing capacity has been fully installed and operationalized.
Q: What is the nature of the new US restrictions mentioned by Tencent management?
A: These restrictions target remote access to advanced US AI processors via computing facilities outside China, meaning they restrict computing services rather than just the physical shipment of chips.