Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

China's NBS confirms 81.8% surge in AI equipment spending as broader investment falls

China's National Bureau of Statistics confirmed Monday that internet companies' equipment spending surged 81.8% year-on-year in the first seven months of 2026, while overall fixed-asset investment fell 6.7%. The data validates the AI infrastructure buildout by Alibaba, Tencent, ByteDance and Baidu, with Tencent's second-quarter operating capex reaching 51.8 billion yuan, up 190% year-on-year.

Published
国家统计局:前7个月互联网企业设备购置大增81.8%,AI算力投资全面落地
Image source: geocountries.com

China's National Bureau of Statistics confirmed Monday that internet companies' equipment spending surged 81.8% year-on-year in the first seven months of 2026, while overall fixed-asset investment fell 6.7% to roughly 26 trillion yuan ($3.85 trillion). It is the most authoritative government measurement yet that China's AI infrastructure buildout has converted from announced commitment into actual hardware purchases.

The divergence is stark sector by sector. Manufacturing investment fell 1.7%, infrastructure investment slipped 3.6%, and mining rose just 3.3% — while the information transmission sector posted fixed-asset investment growth of 29.2% and integrated circuit manufacturing investment rose 11.6%. Even excluding property, overall fixed-asset investment declined 3.7%, making the internet-sector surge the lone exception in a contracting economy.

Corporate earnings had telegraphed the trend. Tencent's second-quarter operating capex reached 51.8 billion yuan ($7.7 billion), up 190% year-on-year and 66% quarter-on-quarter, directed at AI infrastructure for its Hunyuan model, WorkBuddy and CodeBuddy inference operations, and Weixin AI features. The spending pushed Tencent into negative free cash flow of 13.8 billion yuan ($2.0 billion) — its first since 2005 — with net cash falling to 58.2 billion yuan in June from 146.9 billion yuan in March.

Alibaba, ByteDance and Baidu are spending in kind. Alibaba's three-year, 380 billion yuan infrastructure commitment may be exceeded — LatePost reported insiders weighing 480 billion yuan in total spending. ByteDance raised its 2026 capex target from roughly 160 billion yuan to more than 200 billion yuan, about half earmarked for chip procurement. Baidu posted 79% AI cloud revenue growth, with GPU cloud revenue up 184%.

Independent estimates now line up with the official data. TrendForce had projected the four companies' combined 2026 capex would grow more than 80% year-on-year — almost exactly the 81.8% the NBS now reports. NBS statistician Sun Xiao added that high-tech and digital product manufacturing account for roughly half of China's industrial output growth, with 3D printing equipment output up 52.3%, industrial robots up 28.5%, and lithium-ion batteries up 40.2%.

The spending has a physical footprint: data centers require servers, GPUs, precision racks, optical interconnects, battery backup and custom cooling, and surging manufacturing output is the supply chain reading the same signal from the other side of the transaction. A separate state plan to spend roughly 2 trillion yuan over five years on data center infrastructure compounds the private-sector wave.

Scale, however, is not the same as efficiency. U.S. hyperscalers plan roughly $725 billion in 2026 capex versus about $100 billion for China's four largest AI spenders — one-seventh the total. And domestic chips trail Western silicon: Huawei's Ascend 910C is built on a roughly 7-nanometer-class process, and its CloudMatrix 384 cluster consumes about four times the power of Nvidia's NVL72 system.

Watch next: how fast Tencent's cash buffer drains and whether its prepaid compute resale strategy pays off, Alibaba's final capex number, execution of domestic chip purchases alongside the approved Nvidia H200 carve-out, and when this investment cycle starts converting into revenue.

Why it matters

Official statistics confirm China's AI infrastructure spending is expanding far faster than the broader economy, a tailwind for the compute supply chain that also signals sustained cash-flow pressure on internet giants and an intensified push for domestic chip self-reliance.

ChinaNBSAI InfrastructureCapex
Back to realtime news

Nearby Updates

All

08/17, 23:55

Alibaba launches HappyShrimp, an AI music model that generates complete songs from text prompts

Alibaba has launched HappyShrimp, a new AI music model that generates complete songs from text prompts, according to a report carried by finance.biggo.com. The release extends Alibaba's generative AI push from language and image tools into the audio domain.

08/17, 22:46

AI Stocks Rally as Anthropic's 14x Revenue Growth Report Boosts Sentiment

AI-related stocks rallied on August 17, with sentiment lifted by a report showing Anthropic's revenue grew 14-fold. The report was seen as direct evidence of stronger commercialization of frontier AI models, restoring investor confidence in the AI sector.

08/17, 21:58

Nvidia's Huang Declares AI Factories the New Revenue Engine

Nvidia CEO Jensen Huang has cast AI factories as the defining infrastructure of the AI era and the company's new revenue engine, arguing that in the AI economy, compute is revenue. The declaration, carried by Nvidia's official blog and covered by tech media, frames AI factories as full-stack builds spanning chips, packaging, memory, networking, land and power.

08/17, 21:56

Alibaba Unveils Laptop-Runnable Large Model to Counter Meta's AI Push

Alibaba has released a brand-new large language model that runs on laptops, according to Sina Finance, which frames the launch as a direct answer to Meta's recent AI offensive. The model's headline feature is local inference on a personal notebook without cloud APIs, signaling that the large-model race is moving onto consumer devices.