Commentary: China Must Look Beyond AI to Fix Its Deep Macroeconomic Flaws

1. Since early 2025, China has aggressively embraced artificial intelligence, cementing its position as the world’s second AI superpower alongside the U.S. However, this AI boom is not a panacea for the country’s macroeconomic woes, including a five-year real estate slump, deflation, and weak consumer demand. [para. 1][para. 2]

2. The U.S. and China have adopted distinct AI paths: the U.S. relies on technological barriers and capital intensity, while China leans into open-source ecosystems, vast user data, and rapid commercial application. China’s “East Data, West Computing” initiative pairs AI’s energy needs with abundant green electricity in the west, building advantages in algorithmic efficiency and industrial deployment. [para. 3]

3. AI is bolstering growth through fixed-asset investment and exports. Core AI industries are projected to exceed 1% of GDP by 2026, and broad AI-related capex—including computing infrastructure, power grids, and data centers—will reach about 1.2 trillion yuan ($180 billion) this year, contributing roughly 0.3 percentage points to GDP growth. However, this is small relative to the real estate sector, which at its peak accounted for 14.5% of GDP and still hovers around 6%. [para. 4][para. 5]

4. AI-related exports, especially chips and servers, surged 79.1% year-over-year in April 2026, but this is driven primarily by soaring global prices amid supply shortages, not organic volume growth. China remains a massive net importer of high-end chips, and the exorbitant cost of imported components worsens the country’s terms of trade. [para. 6]

5. The AI boom is interacting with the property bust to create two K-shaped divergences. First, demographic: AI replaces routine cognitive tasks, shifting national income from labor to capital. Winners are capital owners and highly educated elites; losers are entry-level white-collar workers. Those displaced by AI are forced into an already saturated gig economy, which now exceeds 200 million people. As wages stagnate and physical AI (autonomous vehicles, robots) accelerates, this employment buffer erodes, suppressing aggregate demand and fueling deflation. [para. 7][para. 8][para. 9]

6. Second, geographic: The AI super-cycle is monopolistic, clustering immense capital, GPUs, and elite talent in just four first-tier hubs: Beijing, Shanghai, Shenzhen, and Hangzhou. These cities monopolize the AI dividend, while inland regions host low-employment, low-value-add data centers under the “East Data, West Computing” scheme. Any meaningful real estate recovery will be confined to these four hubs; for most Chinese cities, a broad property rebound is a mirage. [para. 10][para. 11]

7. Policy implications: Beijing must develop a self-reliant AI supply chain but shed the illusion that AI can solve deep-rooted macro imbalances. It needs to expand the social safety net to catch displaced workers, consider moderating autonomous technologies that threaten livelihoods, and systematically resolve the real estate debt crisis. Drawing lessons from Japan’s NPL disposal (2002-2004), China must enforce strict accounting standards, accelerate bad-debt write-offs, and inject public capital. Overhauling the fiscal system to provide local governments with sustainable tax revenues is also critical. [para. 12][para. 13][para. 14][para. 15]

8. In conclusion, China’s AI boom is real but not an economic cure-all. Only by pairing technological ambition with aggressive debt restructuring and robust social protections can Beijing balance growth with stability in the AI era. [para. 16]