ByteDance Plans Up to $70 Billion in AI Data Center Spending for 2026

ByteDance is planning capital expenditure of up to $70 billion on AI data centers and infrastructure in 2026, according to Bloomberg. The figure represents a near-tripling of the company's 2025 capex of approximately $25 billion, and would position ByteDance alongside Microsoft, Google, and Amazon as one of the largest AI infrastructure spenders in the world.

Self-funded from profits

Unlike many AI investments that rely on external fundraising, ByteDance plans to fund most of the spending from its own cash reserves. The company generated approximately $50 billion in profit in 2025 — making it one of the most profitable tech companies globally and giving it the financial capacity to accelerate without diluting shareholders or taking on debt.

Scale in context

To put $70 billion in perspective: Microsoft committed approximately $80 billion in AI data center spending for its 2025 fiscal year. Google's parent Alphabet disclosed $75 billion in 2026 capex. ByteDance's planned $70 billion would put it directly in the same tier — remarkable for a company that, unlike Google or Microsoft, does not sell cloud services as a primary business.

The spending is concentrated on GPU clusters, networking infrastructure, and data centers needed to train and serve large-scale AI models. ByteDance operates its own internal AI models, including Doubao (its consumer AI assistant popular in China), as well as infrastructure that powers TikTok's recommendation and content moderation systems globally.

Strategic context

ByteDance has been accelerating its AI investments since 2024 amid growing competition from Chinese rivals including Baidu, Alibaba, and the fast-moving DeepSeek. At the same time, TikTok's ongoing regulatory uncertainty in the United States has pushed the company to diversify its AI capabilities and international infrastructure footprint.

The $70 billion figure represents a strategic bet that compute scale will remain a decisive competitive advantage in the next phase of AI development — a bet that the largest US hyperscalers are making simultaneously.