Sources: ByteDance plans to spend ~$5.5B to acquire AI chips in China in 2025, double the amount spent in 2024, with ~60% going to Chinese suppliers like Huawei
Context & Ripple Effects
ByteDance’s planned 2025 purchase budget marks an early, large-scale commitment to AI infrastructure, with the majority directed to domestic suppliers including Huawei. It establishes the procurement base beneath its later preliminary $23B AI-capex plan for 2026.
The subsequent coverage also shows that ByteDance’s chip sourcing was not exclusively local: it later planned substantial Nvidia AI-chip purchases for 2026. That makes the 2025 allocation a concrete example of a split-sourcing approach rather than a simple supplier replacement.
First-order effects
- ByteDance’s planned AI-chip outlay doubles from 2024 to about $5.5B in 2025, immediately increasing its demand for AI compute in China.
- Chinese vendors, including Huawei, are positioned to receive roughly 60% of that budget, giving local suppliers a larger share of ByteDance’s near-term orders.
Second-order effects
- A larger committed buyer can help Chinese chip suppliers plan production, software support, and customer deployments around a major domestic workload.
- The remaining spending creates a procurement contest between domestic and non-Chinese suppliers, while forcing ByteDance to manage a mixed hardware estate.
Third-order effects
- If comparable buyers sustain this pattern, China’s AI infrastructure market could develop around parallel local and foreign hardware supply chains rather than a single dominant stack.
- The later rise in ByteDance’s projected capex suggests chip procurement is becoming a central strategic input; the durability of the domestic share will depend on suppliers meeting deployment needs at scale.
The trend: This is one data point in the AI hardware strategy split, where major Chinese AI buyers expand compute capacity while balancing domestic sourcing with access to foreign accelerators.