Tencent and ByteDance Surpass Meta and Google in NVIDIA AI GPU Purchases

In 2024, two major Chinese tech companies, Tencent and ByteDance, made a big impact in the global AI hardware competition by purchasing a large number of advanced NVIDIA GPUs. Their massive purchases of NVIDIA AI GPUs even outpaced tech giants like Meta and Google, signaling a major shift in the AI hardware market. 

According to reports from Omdia and the Financial Times, both Tencent and ByteDance acquired around 230,000 Hopper GPUs each, including the H20 versions designed to meet the US-China trade restrictions. This put them ahead of major US tech companies like Google, Meta, and Amazon.

However, in this hardware race, Microsoft led the pack with a record-breaking 485,000 Hopper GPUs purchased. Microsoft’s huge investment in AI hardware highlights its ambitious strategy to push for AI innovation, far surpassing the purchases made by Tencent, ByteDance, and other competitors.

While Tencent and ByteDance have made impressive strides in acquiring GPUs, American tech companies are quietly shifting towards developing their own chips. Google has deployed 1.5 million Tensor Processing Units (TPUs), Meta has expanded its use of MTIA chips, and Amazon relies on its own Trainium and Inferentia processors. Although Microsoft still depends heavily on NVIDIA, it has begun testing its Maia chips, with 200,000 units already installed. This shift reflects US companies’ careful focus on technological independence and supply chain security.

The global race between tech giants is far from over. Microsoft’s enormous investments in NVIDIA GPUs clearly show its aggressive AI strategy, while Tencent and ByteDance’s strategic purchases highlight their determination to stay competitive, even under geopolitical pressures. These moves are not just about showcasing technological power but also about securing their place in the future AI ecosystem.

Looking ahead to 2025, the focus will be on how these companies navigate trade restrictions and lead the next wave of AI innovation. Despite facing trade barriers, Tencent and ByteDance managed to surpass Meta and Google in acquiring NVIDIA GPUs, demonstrating China’s strong commitment to the AI field. However, their heavy reliance on NVIDIA presents a potential risk. If export restrictions tighten further, depending on a single supplier could limit their growth.

In contrast, US companies like Microsoft, Google, and Amazon are reducing external dependencies by developing their own chips. This strategy helps them control costs, optimize performance, and enhance technological security. By using proprietary processors like TPUs and MTIA chips, companies like Google and Meta show foresight in reducing dependence on outside suppliers.

NVIDIA’s dominance in the AI server hardware market, holding 43% of the market share in 2024, demonstrates its strong leadership in AI solutions. However, as competition intensifies in the AI hardware space, the future remains uncertain. In this quiet battle, companies that can balance innovation with supply chain security will likely have a better chance of standing out in the global AI race.