China’s economic divergence deepens: technology and AI support the long-term transition, while domestic demand and real estate remain weak

Overview

The report discusses the further divergence in China’s economy during the second quarter of 2026: exports, industrial production, high-tech manufacturing, AI, and data center-related investment remain relatively strong, while consumer spending, real estate, domestic auto sales, and fixed asset investment are weak. Goldman Sachs argues that the long-term technology transition remains the policy centerpiece, but if cyclical policy is insufficient, weak employment, income, and confidence could further undermine domestic demand and, in turn, slow the formation of a technology-driven growth model.

Core views

First, the April-May data point to slowing economic momentum. Exports contributed roughly 3 percentage points to real GDP growth, implying that domestic demand growth under the official data was only about 1%-2% year over year. Second, energy price shocks, slower fiscal spending, and unfavorable weather were important reasons for the second-quarter weakness; as a result, Goldman Sachs cut its second-quarter growth forecast, but expects lower oil prices, faster fiscal spending, and weather normalization to drive a rebound in the third quarter. Third, China’s economic structure is shifting from property-driven to technology-driven, with industrial robot and semiconductor output rising significantly versus 2019, while output in glass, cement, and other property-related industries has fallen. Fourth, AI diffusion is both a long-term productivity opportunity and a source of near-term macro headwinds through job substitution, income pressure, and downward pressure on rents.

Analysis framework

The report combines official macro data, implied demand from exports and imports, sector output, fiscal spending, local government special bond issuance, sector performance in equity markets, data center capacity, and AI capex comparisons to assess China’s cyclical momentum and structural transition. The analysis focuses not only on the headline GDP target, but also on how employment, consumption, real estate, fiscal pacing, and confidence affect whether the long-term technology transition can proceed smoothly.

Methodology notes

• macro_forecastReal GDP forecast revision

  Adjust quarterly annualized quarter-on-quarter growth forecasts based on the latest April and May activity data while keeping the full-year forecast unchanged.

  The second-quarter forecast was cut from 4.0% to 3.5%, while the third-quarter forecast was raised from 4.5% to 5.0%, reflecting a shift between short-term shocks and subsequent recovery.

• growth_decompositionExport contribution and domestic demand decomposition

  Estimate real domestic demand growth using exports’ contribution to GDP growth.

  The report estimates that export growth contributed about 3 percentage points to real GDP growth year over year, implying domestic demand growth of only about 1%-2% when overall growth is around 4%.

• structural_transitionTechnology-driven transition framework

  Compare the divergence between high-tech manufacturing, AI, and data centers on one side and real estate, consumption, and traditional building materials on the other.

  This framework is used to show that the policy goal is to shift the economy away from property dependence and toward technology and AI, but weak cyclical domestic demand may affect the quality of that transition.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• China macro assets

  core research object

  Strengths

  Exports remain strong, industrial production is still resilient, and long-term policy planning is clear.

  Weaknesses

  Domestic demand growth is weak, and consumption, real estate, and fixed asset investment are under pressure.

  Comparison

  The export sector is clearly stronger than domestic demand sectors, creating a disconnect between aggregate growth and micro-level experience.

  Risks

  If cyclical policy support is insufficient, weak confidence and employment may weigh on the long-term transition.

• China technology and AI industry chain

  structural beneficiary

  Strengths

  High-tech manufacturing, information transmission, R&D, and AI-related investment are growing much faster than the overall economy.

  Weaknesses

  Capex statistics are not directly comparable with those in the United States, and AI diffusion may create job substitution pressure.

  Comparison

  In the United States, AI investment is driven more by hyperscale cloud providers such as Amazon, Meta, Google, Microsoft, and Oracle, while in China it relies more on lower construction costs and government investment.

  Risks

  If AI adoption substitutes for labor too quickly, it could suppress income, consumption, and the recovery in real estate.

• China consumer and real estate

  macro drag

  Strengths

  After years of decline, housing prices may be nearing a bottom over the next one to two years.

  Weaknesses

  Retail sales, domestic auto sales, rents, and household confidence remain weak, and employment pressure has not yet eased.

  Comparison

  Relative to the technology sector, consumption and the property chain lag materially in both physical output and equity market performance.

  Risks

  Youth unemployment and AI substitution could lead to stagnant or even falling income, delaying recovery in real estate and consumption.

• China fiscal and policy tools

  short-term stabilizer

  Strengths

  Government bond issuance and fiscal spending still have room to accelerate, and cyclical policy can buffer external and energy shocks.

  Weaknesses

  There has historically been a pattern of fiscal slowing after strong quarters and re-acceleration after weak data.

  Comparison

  Long-term planning is a clear advantage, but insufficient short-term policy consistency may weaken confidence.

  Risks

  If policy focuses only on aggregate GDP rather than employment and consumption, it may limit domestic demand recovery.

Key data

• 2026 full-year real GDP forecast4.7%Goldman Sachs kept the full-year forecast unchanged.
• 2026 second-quarter real GDP annualized quarter-on-quarter forecast3.5%Cut from 4.0%; the year-over-year forecast was lowered from 4.7% to 4.5%.
• 2026 third-quarter real GDP annualized quarter-on-quarter forecast5.0%Raised from 4.5%, assuming lower oil prices, faster fiscal spending, and weather normalization.
• May retail sales-0.6% yoyShows weakness on the consumption side.
• Fixed asset investment for January-May-4.1% yoyA rare historical instance in the official data where both retail sales and fixed asset investment were negative.
• Export growthclose to 20% yoyIn nominal U.S. dollar terms, still the strongest part of the economy.
• May domestic auto sales and exportsdomestic sales -22% yoy, exports +75% yoyShows a pronounced divergence between external demand and domestic demand.
• May value added in high-tech manufacturing+15.1% yoyFar above overall industrial value added growth of 4.5%.
• China data center installed capacity60% of the U.S.As of mid-2025, indicating that China’s AI compute base is not measured solely by hyperscaler capex.
• 15th Five-Year compute network investment planRMB2tnIncludes data center and other compute-network construction.

Impact & implications

For investors, the report reinforces the view that there is a structural divergence within China’s asset universe: technology, AI, semiconductors, data centers, and high-tech manufacturing continue to receive policy and capital support, while consumption, real estate, and parts of the traditional cyclical chain remain constrained by employment, income, confidence, and fiscal pacing. At the macro level, simply pursuing the full-year GDP target may not be enough to repair domestic demand; if policy shifts more toward employment and consumption objectives, it would be more conducive to a virtuous cycle between technology-led productivity gains and domestic demand expansion.

Risks

• Rapid AI diffusion leading to job substitution for white-collar and entry-level positions, increasing income and consumption pressure.
• If housing prices and rents continue to fall due to weakening employment, the bottoming process in the property market could be delayed.
• Strong exports may mask weak domestic demand; if policy focuses only on aggregate GDP, domestic demand growth could remain constrained.
• Frequent quarter-to-quarter swings in fiscal spending and policy stance may undermine household and corporate confidence.
• Energy price shocks, extreme weather, and El Niño-related climate anomalies could continue to disrupt production, consumption, and investment.
• U.S.-China competition, technology transfer regulations, and cross-border investment restrictions may affect technology supply-chain allocation.

What to watch

• Whether industrial production, retail sales, fixed asset investment, and auto sales recover in June and throughout the third quarter.
• Whether local government special bond issuance, fiscal spending, and policy-finance tools accelerate.
• Employment data, youth unemployment, entry-level hiring, and household income trends.
• Whether property prices, rent inflation, and transaction volumes can approach a bottom.
• Growth in high-tech manufacturing, semiconductors, industrial robots, AI computing power, and data center investment.
• Whether policy objectives shift from a single GDP growth target toward employment, consumption, and domestic demand.
• The impact of oil prices, energy supply, and weather conditions on quarterly growth momentum.