Chinese factory activity returns to growth in September amid AI boom

China's official manufacturing PMI rose to 50.1 in September from 49.8 in August, ending two months of contraction and matching consensus, while non-manufacturing PMI rebounded to 50.2 from 49.0. Production reached 51.7 and new orders 50.5, aided by the global AI boom and reduced weather disruption, but weak retail sales, investment and a prolonged property downturn continue to constrain domestic demand. Beijing has announced cheaper-credit support for infrastructure, technology and homebuyers, while China and the U.S. will pursue tariff reductions covering $60 billion of bilateral imports despite notable exclusions including non-seed soybeans.

Analysis

The key read-through is not a broad China-demand inflection but a narrow industrial stabilization led by production capacity and external demand. Production running ahead of new orders raises the probability of renewed inventory accumulation and price competition over the next 1-3 months, limiting margin upside for Chinese manufacturers even if volume data improve. This favors upstream inputs and policy-linked infrastructure beneficiaries over export-oriented finished-goods producers, whose incremental output faces tariff and anti-dumping risk.

The credit-support impulse is potentially more relevant for state-linked banks, grid investment, rail, and construction machinery than for discretionary consumption. A durable rerating in China equities requires evidence that household credit demand and property transactions respond; cheaper funding alone can worsen bank net-interest-margin pressure if loan demand remains weak. For global cyclicals, the more important transmission channel is incremental metals and power-grid demand, not a near-term revival in Chinese consumer imports.

Consensus may over-read a move above the 50 PMI threshold while underweighting the divergence between industrial resilience and weak domestic absorption. The immediate equity reaction should therefore be modest, particularly while higher U.S. real yields constrain EM multiples. Over 6-18 months, an export-heavy adjustment path increases the likelihood of trade barriers, which would shift supply chains toward Mexico, ASEAN, and India rather than generate a clean China-led global growth upswing.

Key Decisions for Investors

• Use a 1-3 month tactical pair: long FXI / short KWEB. FXI has greater exposure to banks, SOEs, energy, and policy-sensitive industrial activity, while KWEB needs a stronger household-demand recovery to outperform. Falsify if nationwide property-sales data improve materially for two consecutive months or consumer-credit growth accelerates ahead of infrastructure lending.
• Maintain a watch-to-buy on FCX or COPX rather than chase immediately. Initiate only if Chinese credit and fixed-asset-investment data confirm a broadened infrastructure impulse; target a 6-12 month copper-demand recovery, with downside protection via a stop if copper breaks below its pre-data range or Chinese construction activity rolls over again.
• Avoid adding to Chinese export-manufacturing and appliance exposures despite the industrial improvement. A tariff-cut framework that excludes major agricultural categories leaves meaningful bilateral-policy risk unresolved; new U.S./EU trade investigations would compress export volumes and force discounting within 3-6 months.
• For global macro books, retain a relative preference for U.S. quality and avoid broad EM-beta additions until long-end Treasury yields stabilize. A China industrial uptick is unlikely by itself to offset multiple compression from rising global discount rates; reassess if U.S. 10-year real yields retreat and Chinese domestic-demand indicators broaden.