China’s AI Boom Faces A Demand Test As Chip Bottlenecks Raise The Cost Of Growth
China’s AI investment exceeds one trillion yuan a year, but Nomura economist Lu Ting says chip constraints, capital-market gaps and city-level concentration limit its wider economic lift.

AI Spending Is Not Spreading Through China Like A Property Boom
China’s artificial intelligence build-out is large, but the economic transmission is narrower than the investment headline suggests.
Lu Ting, chief China economist at Nomura, said at a media briefing in Beijing on Thursday that AI is not lifting China’s economy as strongly as it is lifting the United States, and that the technology could also widen domestic economic divides.
The scale is still material.
Annual AI investment in China roughly exceeds one trillion yuan, or US$147.82 billion, per year.
Lu’s point is that this remains below the force once generated by property investment during the 2010s, while AI’s share of China’s overall economy is only about one-third of the US ratio.
That makes AI a different kind of growth engine for China.
Real estate and new energy investment sent demand through many cities, suppliers and households.
AI development is more concentrated, with Lu identifying Beijing, Shanghai and Hangzhou as the kind of large-city hubs that can absorb the sector’s high entry barriers.
Chips And Capital Markets Define The Constraint
The comparison with the US is partly about financing and partly about access to hardware.
Lu said AI drives about half of the US economy, while US AI investment has expanded at roughly four times the pace of consumer spending.
He also linked that strength to developed capital markets, where companies such as OpenAI can raise money more easily.
China faces a different bottleneck.
Lu said bulk chip buying is limited because other countries will not sell the needed products.
That constraint matters because AI gains are also flowing to economies with semiconductor production and exports.
South Korea and Taiwan were identified as major beneficiaries of the AI boom, and Taiwan’s first-quarter GDP growth reached 14.55 per cent, compared with mainland China’s 5 per cent growth in the same quarter.
The World Bank’s latest China outlook adds another check on the growth story.
On Thursday, it forecast full-year growth of 4.2 per cent for China, down 0.2 percentage points from its January forecast.
Export Strength Does Not Remove The Import Bill
China is still a major exporter of legacy chips, but the AI trade effect is not cleanly positive.
In May, China exported US$35.5 billion worth of integrated circuits, up 110.9 per cent year on year, while export volume rose only 2.1 per cent.
Import values reached US$56.6 billion, up 68 per cent, even as import volume fell 1 per cent.
Those figures show why AI-related trade can raise costs for domestic firms even when some exporters benefit.
Lu said some Chinese companies are winning from the AI export angle, but many more are paying higher prices for the same items, limiting the overall contribution to GDP.
The Watchpoint Is Domestic Demand, Not Only Model Capability
The source material does not show that China lacks AI companies or AI investment.
It shows a distribution problem: capital, chips, skilled labour and platform effects are concentrated in a smaller set of firms and cities.
Lu warned that large language model providers in top-tier hubs could replace lower-end white-collar professionals in smaller and medium-sized cities without creating the positive outward spillover associated with earlier booms.
The property backdrop sharpens that risk.
Shanghai is showing early signs of recovery in secondary real estate sales and prices, while many smaller cities remain stuck in a downturn.
For SendTech Times readers, the key test is whether China’s AI investment begins to support broader domestic demand, or whether it remains a high-capex sector whose benefits cluster around leading cities, semiconductor importers and firms with privileged access to computing resources.
















