Tencent admits GPUs only pay for themselves when powering personalized ads

Tencent has revealed that its investment in graphics processing units (GPUs) only becomes financially viable when used to power its personalized advertising business. During its Q1 2026 earnings call, Chief Strategy Officer James Mitchell explained that deploying GPUs in ad technology leads to better targeting, higher click-through rates, and accelerated revenue and profit growth. However, the company struggles to generate immediate returns from GPUs used for its AI projects, such as the Hunyuan foundation model, which it views as a long-term strategic investment. The company also faces challenges in GPU procurement due to US sanctions and limited semiconductor fabrication capacity within China. Chief Financial Officer Shek Hon Lo noted that these issues are gradually being addressed as Chinese-designed GPUs and ASICs become more available from domestic and neighboring country fabs. Despite this progress, Tencent expects GPU acquisition to remain more difficult than CPU procurement, where it maintains longstanding relationships with major suppliers like Intel and AMD. Tencent’s GPU shortage has constrained its ability to rent accelerators through its public cloud services, limiting its competitiveness in that market. The company anticipates that increased supply of locally designed GPUs will help alleviate this bottleneck. Meanwhile, Tencent reported strong overall financial performance, with revenue reaching RMB196.5 billion ($28.9 billion), a 12 percent increase. Its messaging platforms Weixin and QQ now serve 1.95 billion monthly users combined, and Tencent has integrated AI capabilities into these apps to enable users to control AI agents across mobile, PC, and cloud environments. This approach contrasts with Western tech giants like Google and Meta, which have yet to develop similar AI-driven communication interfaces and do not face the same hardware supply constraints. Tencent’s experience highlights the complex interplay between geopolitical factors, supply chain limitations, and the evolving economics of AI infrastructure investment in China’s tech sector.