Tencent self-sufficient in AI chips, dampening US sales hopes
Chinese technology behemoth Tencent has declared its independence from American graphics processing unit (GPU) imports, asserting that it already possesses a sufficient supply of chips to meet its needs. This surprising announcement, made during the company’s second-quarter earnings call, casts a shadow over the hopes of US chipmakers like Nvidia and AMD, who anticipated a surge in revenue following Washington’s recent decision to permit renewed GPU sales to China.
When questioned by a financial analyst about the impact of the US policy shift, Tencent President Martin Lau stated, “We do not have a definite answer on the import situation yet. There are a lot of discussions between the two governments.” He quickly added, however, “From our own perspective, we do have enough chips for training and continuous upgrade of our existing models. We also have many options for inference chips.” Lau further revealed that Tencent is actively implementing significant software improvements to enhance the efficiency of its AI inference operations, allowing more workloads to be processed on the same hardware footprint.
Tencent’s confident stance delivers a fourfold blow to various stakeholders. Firstly, it signals a potential dampening of ambitions for AMD and Nvidia, who were poised to capitalize on renewed access to the vast Chinese market. Secondly, it complicates the calculus for the US government, particularly the Trump administration, which reportedly plans to levy a share of future GPU sales to China. Thirdly, Tencent’s casual mention of “many options for inference chips” subtly suggests it can source these crucial components from non-US manufacturers, further eroding the leverage of American tech firms. Lastly, the company’s emphasis on software optimization indicates a strategic shift towards extracting greater performance from existing hardware, potentially leading to a long-term deceleration in chip procurement.
Beyond the immediate implications for hardware sales, Lau also tempered investor enthusiasm regarding the profitability of AI. When asked about the impact of AI investments on Tencent’s margins, he acknowledged, “Depreciation cost from AI will continue to go up.” While he affirmed that Tencent continues to benefit from AI, he cautioned, “The issue is these two may not match each other completely, but both are moving in the same general direction.” This candid assessment suggests that Tencent may be facing challenges in fully monetizing its AI services and finding sufficient paying customers to offset the escalating costs associated with AI infrastructure.
This strategic pivot is also evident in Tencent’s cloud business, where Lau stated the company is pursuing opportunities “not dependent on the vagaries of the GPU supply situation.” He explicitly clarified, “Our cloud strategy is not dependent on GPU. We are also growing in CPU and database.” This marks the third consecutive quarter that Tencent has conveyed to investors that it does not require additional GPU purchases, underscoring a consistent and deliberate strategy to reduce reliance on external hardware supply chains.
Despite these cautious remarks on AI profitability and chip procurement, Tencent remains in robust financial health. The company reported a second-quarter revenue of RMB184.5 billion ($25.7 billion), marking a 15 percent year-over-year increase. Net profit also saw healthy growth, rising 11 percent to RMB 64.8 billion ($9 billion). Furthermore, the average monthly active users of its flagship Weixin and WeChat social networks climbed to 1.411 billion, reflecting a 3 percent annual increase of 40 million users. Tencent’s position, therefore, is not one of distress but rather a calculated shift towards self-sufficiency and optimized resource utilization in a volatile geopolitical and technological landscape.