China’s $70 Billion AI Data-Center Boom

Archive edition · Market data and company circumstances reflect 8 November 2025, when this newsletter was sent.

China’s AI Power Play

China is quietly engineering one of the biggest infrastructure expansions in its history: A nationwide push to build the digital backbone for artificial intelligence.

What started as a breakout moment for a single chatbot has now evolved into a multi-billion-dollar race involving the country’s largest tech giants, energy providers, and policymakers.

This newsletter explores how that transformation is unfolding; from data-center growth and domestic chipmaking to global expansion and early signs of monetization.

The Birth of China’s AI Momentum

The spark came in January 2025 when DeepSeek, a home-grown open-source chatbot, became China’s most-downloaded app within weeks. The phenomenon mirrored ChatGPT’s U.S. debut, igniting national pride and a sense of technological urgency.

Suddenly, China’s internet giants, Alibaba (BABA), Tencent (0700.HK), Baidu (BIDU), and Huawei; rushed to scale cloud infrastructure, train large-language models, and develop domestic AI chips.

These firms will invest over $70 billion in 2026 to expand data centers and AI capabilities. Analysts expect power demand from AI computing to jump 25 % this year alone, pushing total data-center capacity toward 30 gigawatts (GW), roughly the energy footprint of a mid-sized European nation.

For Beijing, the DeepSeek moment validated years of policy promoting “indigenous innovation.” For investors, it signaled that China’s AI sector had entered its hyper-growth phase, where infrastructure spending precedes monetization.

The Infrastructure Race

AI’s engine is electricity. Every training run, every inference request, every generated video or image consumes massive compute power. Chinese hyperscalers are responding with a historic build-out of high-density data centers and chip foundries.

Local governments, from Guizhou to Inner Mongolia, are offering renewable-energy contracts and tax incentives to attract these facilities, creating inland “AI compute hubs.”

Meanwhile, China’s “East Data, West Compute” strategy channels excess hydropower and solar capacity into data-center corridors. It is not just about AI, it is about energy security, regional development, and digital sovereignty.

• Electricity capacity: Expected to rise 30 % YoY to ~30 GW.
• Capex growth: Cloud providers will lift investments ~65 % in 2025.
• Ten-year goal: One major player plans a 10× capacity increase by 2032.

Domestic Self-Reliance & Emerging Winners

Historically, Chinese technology firms allocated 50–75% of their AI capital expenditure to foreign chips, particularly Nvidia GPUs, which powered most high-end training clusters. However, recent U.S. export restrictions have accelerated Beijing’s long-standing goal of technological self-sufficiency, forcing hyperscalers to localize their supply chains.

Leading the transition is Huawei Technologies Co., whose Ascend 910B processors have emerged as China’s flagship AI chips, now powering next-generation servers across public and private cloud platforms.

Baidu Inc. (BIDU) is leveraging its in-house Kunlun series to train and deploy its ERNIE large-language model ecosystem, while Cambricon Technologies (688256.SS) has positioned itself as a key domestic designer of AI accelerators and edge processors.

On the semiconductor front, Semiconductor Manufacturing International Corp. (SMIC – 0981.HK) continues to expand its advanced fabrication nodes, ensuring local access to compute-grade wafers previously sourced from abroad.

Complementing this hardware base, Inspur Information (000977.SZ) is developing AI-optimized server architectures that integrate seamlessly with home-grown chips, closing the loop on China’s hardware dependency.

At the cloud and application layer, Alibaba Group Holdings (BABA) is scaling AliCloud’s AI training clusters across Asia, while Tencent Holdings Ltd. (0700.HK) focuses on enterprise AI solutions and generative-content platforms.

Together, these companies form the pillars of a vertically integrated AI ecosystem designed to rival U.S. hyperscalers on both performance and cost. With strong policy support and state-backed funding visibility, Beijing’s localization drive is rapidly transforming what was once an import-reliant sector into a strategically self-sufficient AI powerhouse.

Outlook to 2030: AI as a National Asset Class

By 2030, China’s AI economy could account for up to 7 % of GDP, rivaling today’s internet sector. The $70 billion spending wave of 2026 is the foundation for a multi-trillion-yuan ecosystem spanning chips, cloud services, and applications.

Short-term profitability will lag, but history suggests compounding returns once utilization rises above 60 %. As with early AWS (AMZN) and Azure (MSFT), scale will eventually flip margins positive.

Strategically, exporting AI infrastructure extends China’s digital influence across emerging markets, giving Beijing leverage in technology standards and data governance.

For investors, key exposure buckets are:

In essence, China’s AI build-out is in its “build it and they will come” stage—high capex today for outsized returns tomorrow. The DeepSeek moment proved domestic demand exists; now comes the decade-long mission to monetize it globally.

• Infrastructure Leaders: BABA, 0700.HK, BIDU, Huawei.
• Semiconductor & Hardware Suppliers: 0981.HK, 688256.SS, 000977.SZ.
• Renewable & Cooling Ecosystem: energy utilities and thermal-management OEMs.

Archive note

This article preserves the analysis in our weekly newsletter sent 8 November 2025. Market prices, forecasts and company circumstances reflect the time of publication and may have changed.

This material is general information, not personal financial advice or a recommendation to trade. Investing and trading involve risk, including loss of capital.