appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Howard Liu

The massive gap in artificial intelligence spending between US and Chinese tech titans may not buy the advantage expected for American giants, as lower domestic costs and heavy state support allow Chinese firms to secure far more computing power per dollar, according to a new report by Moody’s Ratings.
While US hyperscalers outspent their Chinese counterparts by a staggering margin, the physical gap in computing capacity was nowhere near as wide as those mega-budgets suggested, the report said.
Lower buildout costs, targeted policy incentives and access to cheaper green energy meant Chinese tech firms were punching above their financial weight, narrowing the compute divide with American peers at a fraction of the price, Moody’s noted.
The US retained a clear overall lead in cutting-edge semiconductor chips, but the findings suggested headline capex numbers tell only part of the story. Here is what you need to know.
How wide is the spending gap between US and Chinese tech giants?
Capital expenditure by China’s major tech companies was set to more than double to around US$140 billion this year – up from US$65 billion in 2025 – and reach US$165 billion by 2027, according to Moody’s.
By contrast, five major US hyperscalers – Microsoft, Amazon Web Services, Alphabet, Meta Platforms and Oracle – plus AI cloud provider CoreWeave were projected to spend more than US$785 billion this year alone, approaching US$1 trillion by 2027.

Why does China’s AI infrastructure cost so much less?
China’s cost advantage lies in land, power and structural state backing.
Under Beijing’s “East Data, West Computing” initiative, power-intensive workloads are routed to resource-rich inland regions like Inner Mongolia and the northwestern province of Gansu.
By moving latency-tolerant AI training to hubs with abundant green energy and cooler climates – such as ByteDance’s expanding data centre footprint in Ulanqab, Inner Mongolia – companies could slash power and cooling expenses while reserving valuable coastal capacity for real-time applications, according to BMI, a Fitch Solutions company.
Government policy further tilted the scales through cheaper infrastructure, subsidies, fast-tracked approvals, and tax incentives, according to Moody’s.
What is holding China back from overtaking the US?
China’s cost advantage is concentrated mainly in non-chip infrastructure.
Servers and accelerators accounted for the largest share of AI spending, while restricted access to leading-edge Nvidia chips remained a key constraint, Moody’s said. Domestic processors and software ecosystems continued to improve but still lagged Nvidia’s.
There is also an efficiency trade-off. BMI said some Chinese semiconductor platforms consumed more electricity than Nvidia processors to perform equivalent computing workloads, meaning cheaper power and subsidies could only partly offset the higher operating costs of domestic hardware.
At the system level, the trade-off is also stark. Semiconductor research firm SemiAnalysis estimated that Huawei Technologies’ CloudMatrix 384 drew about 4.1 times as much power as Nvidia’s GB200 NVL72, while delivering nearly twice the computing performance, making access to abundant and relatively cheap electricity particularly important.
Is China actually closing the compute gap?
Measured by data centre capacity rather than pure dollar spend, the divide shrinks significantly.
Installed US data centre capacity stood at 52 gigawatts (GW) at the end of 2025, compared with 28GW in China, according to International Energy Agency figures cited by Moody’s.
By 2030, US capacity was forecast to reach 100GW, compared with 67GW in China. China’s capacity was expected to grow at an annual rate of 19 per cent, faster than the US rate of 14 per cent, narrowing the relative gap even as the absolute difference increased.




While megawatt output does not directly equate to AI processing power, it shows that the capex gap overstates the physical capacity divide. China’s faster growth rate means it is steadily narrowing the relative gap even as the absolute difference expands.
Model capabilities reflect this catch-up. Leading Chinese models – including Moonshot AI’s Kimi K3, Z.ai’s GLM-5.3, and Alibaba Group Holding’s Qwen3.8 – now rank alongside top global open-weight rivals on benchmark platforms like Artificial Analysis and Chatbot Arena.
Can Chinese tech giants turn these AI gains into revenue?
Building compute is one thing; monetising it is another.
Cloud platforms remained the main engine for AI revenue, but Chinese hyperscalers operated at a much smaller scale and in a more price-competitive market, Moody’s said.
In their latest reported quarters, Microsoft generated US$59 billion in cloud revenue, Amazon US$42 billion, and Alphabet US$25 billion. Alibaba generated US$7 billion and Baidu about US$1 billion.
US hyperscalers also benefited from larger enterprise customer bases, contracted cloud commitments and integrated software ecosystems, giving them clearer paths to recurring AI revenue. China’s lower-cost open-source models were driving adoption, but monetisation remained at an earlier stage, Moody’s said.
There are signs the gap is beginning to narrow. Alibaba’s cloud business reported 45 per cent revenue growth in the June quarter – the fastest pace in 22 quarters – as demand for AI-related products accelerated.
Alibaba owns the South China Morning Post.


Huawei unveils latest tech to boost AI power in push to break China’s Nvidia reliance

Hong Kong’s 5-year plan, policy continuity offer investors more stability, experts say

Huawei accused of stealing robot tech to get hands on T-Mobile’s trade secrets

Hong Kong aims to raise innovation spending to 3% of GDP by about 2030

Artificial intelligence (AI)

Punching above their weight: how China’s AI giants stretch each dollar in compute race

US hyperscalers outspend their Chinese counterparts by a staggering margin, according to a report. Photo: Shutterstock

An employee works at a semiconductor factory in eastern China’s Jiangsu province. Photo: AFP

Frontier AI at a cost: what Anthropic’s Fable 5.1 means for the US-China model race

Anthropic, based in San Francisco, specialises in developing advanced AI models. Photo: Reuters

Tech race flashpoint: Huang, Musk push back on AI guard rails amid heated US-China rivalry

Jensen Huang speaks at the G20 Innovation Ministerial summit. Photo: Reuters

Elon Musk speaks via video conference to delegates at the G20 Innovation Ministerial summit. Photo: Reuters

How China’s AI race is shrinking job market, forcing workers into low-paid, manual roles

A woman holds an umbrella as she walks through Beijing’s Central Business District (CBD). Photo: Reuters

ByteDance to expand massive AI data centre cluster in Inner Mongolia, source says

The ByteDance logo is seen at the company’s office building in Shanghai. Photo: Reuters

The Jining district of Ulanqab in Inner Mongolia autonomous region. Photo: Xinhua

Tech

Big Tech

Tech Trends

Tech War

Business

China Business

Companies

News

China Future Tech

Lower costs and faster growth rates let China’s AI firms narrow the divide, raising the question: does the spending gap still matter?


Punching above their weight: how China’s AI giants stretch each dollar in compute race

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Lower costs and faster growth rates let China’s AI firms narrow the divide, raising the question: does the spending gap still matter?

Lower costs and faster growth rates let China’s AI firms narrow the divide, raising the question: does the spending gap still matter?