Goldman Sachs: China's AI infrastructure narrative reignites, enterprise-level large model token consumption surges, Alibaba is the key beneficiary

```

With the surge in enterprise-level application demand, the investment narrative surrounding China’s AI infrastructure is being reignited.

According to a recent report by Goldman Sachs, full-stack cloud providers, represented by Alibaba, are embracing new growth opportunities, thanks to accelerated enterprise adoption of large models and persistent strong computing demand.

Data shows the actual deployment and application of AI in the commercial sector is significantly speeding up. The average daily Token consumption of enterprise-level large models in China reached 10.2 trillion in the first half of 2025, a sharp increase of 363% compared with the second half of 2024.

Goldman Sachs believes Alibaba, with its leading model capabilities, 47% share of the Chinese public cloud market, and diversified chip supply, is in a strong position and has room for international expansion.

Based on an optimistic view of the industry outlook, Goldman Sachs raised Alibaba’s target price from $163 to $179, maintaining a “Buy” rating. They especially increased their valuation of Alibaba Cloud from $36/ADS to $43/ADS, while raising growth expectations for Alibaba Cloud’s fiscal 2026 Q2–Q4.

Goldman Sachs analysts Ronald Keung et al. emphasized that Chinese cloud providers have made progress in self-developed inference chips and adopted a “multi-chip strategy,” which means the growth of China’s AI cloud industry “no longer depends solely on overseas chip supply.” Combined with strong capital expenditure prospects, this shift brings compound growth potential to the industry.

Enterprise Adoption Rate Soars, Token Consumption Skyrockets

Chinese enterprises are embracing generative AI at an unprecedented pace, which is the core driving force behind the infrastructure narrative. According to Frost & Sullivan, in the first half of 2025, the average daily Token consumption of China’s enterprise-level large models hit 10.2 trillion, a surge of 363% compared with the previous period.

Among many model providers, Alibaba, ByteDance, and DeepSeek have become the top three choices for Chinese enterprises when selecting general-purpose large models.

Citing an Omdia report, Goldman Sachs notes that Alibaba has successfully penetrated the majority of Chinese Fortune 500 enterprises deploying generative AI by 2025, ranking first in this segment. This business model, which bills via API calls and Token usage, places large cloud service providers with leading AI models and agent capabilities in the best position.

Infrastructure Narrative Rekindled, Multi-Chip Strategy Breaks Bottleneck

Goldman Sachs believes the investment narrative for China’s AI infrastructure has been reignited after Alibaba announced better-than-expected cloud business revenue and capex. They forecast that Chinese cloud service providers’ (CSPs) capex will increase 39% year-on-year in Q3 2025, providing a hardware foundation for sustained AI cloud revenue growth.

Meanwhile, the evolving chip supply landscape has also boosted market confidence. Goldman Sachs believes that this diversified chip supply strategy is reshaping the outlook for China’s AI cloud industry.

Based on these positive trends, Goldman Sachs reiterates its “Buy” rating on Alibaba and raises its 12-month SOTP target price to $179. They raised Alibaba Cloud’s fiscal 2026 Q2–Q4 growth forecast from 28%-30% to 30%-32% to reflect its latest full-stack AI products, robust computing demand, and capital expenditure prospects.

The report likens China’s current market stage to that of the US two years ago (end of 2022 when ChatGPT launched), when the market rewarded companies that “built solid infrastructure believing customers would follow” and increased capital spending.

Goldman Sachs analysts believe Alibaba, with its leading model capabilities, 47% share of the Chinese public cloud market (according to IDC), and diversified chip supply, is in a strong position and has room for global expansion. Investors are closely watching the upcoming Yunqi Cloud Summit in Hangzhou, looking for updates on Alibaba Cloud and Alibaba’s AI progress.

Giants Start an Arms Race, but Monetization Still Has a Long Way to Go

Beyond the infrastructure layer, China’s AI model and application layers are also developing rapidly. The report highlights several key advances:

Alibaba: On September 12, released the next-generation model architecture Qwen3-Next, claiming 10x the performance of previous models at only 1/10th the build cost. The Qwen3-Next-80B-A3B model, with 80 billion parameters, runs 10 times faster than the 32B model released in April.

Baidu: On September 9, released Wenxin (ERNIE) X1.1, which saw significant improvements in authenticity, instruction-following, and agent abilities, achieving performance on par with GPT-5 and Gemini 2.5 Pro.

Tencent: Released leading 3D world generation model, HunyuanWorld-Voyager, and high-resolution text-to-image model, HunyuanImage 2.1.

At the same time, AI is starting to be embedded as an “Agent” in mainstream applications. For example, Meituan launched “Xiaomei,” an AI assistant based on its self-developed large model that supports voice ordering and restaurant reservations; Alibaba’s Amap (Gaode Map) launched “Mr. Gao,” an AI-native guide aimed at transforming map services into a personalized travel companion.

Goldman Sachs states that in multimodal areas such as text-to-video and text-to-image, Chinese models are rapidly narrowing the gap with global peers. However, in terms of business monetization, Chinese AI applications still have a long way to go.

By August 2025, the total annual recurring revenue (ARR) of global AI applications is about $30 billion, while the ARR of Chinese AI applications is only $1.5 billion, accounting for 5% of the global share.

This article is from the WeChat public account “Hard AI”. For more cutting-edge AI news, click here.

Risk warning and disclaimerThe market involves risk, and investments should be made cautiously. This article does not constitute personal investment advice and does not take into account individual users’ particular investment goals, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article apply to their particular circumstances. Investment is at your own risk. ```