China factory activity slips in May as economic momentum softens By Investing.com

China’s official manufacturing PMI slipped to 50.0 in May from 50.3, signaling stalled factory growth, while the non-manufacturing PMI improved to 50.1 from 49.4 and returned to expansion. The data point to slowing momentum after a strong start to the year, even as Beijing has responded with policy easing, including a record-low one-year policy loan rate from the PBoC. Export resilience tied to AI-related demand remains a support, but stronger yuan pressure and softer domestic activity are headwinds.

Analysis

The setup is less about “China softening” and more about a split economy: domestic cyclicals are decelerating while export-heavy hardware tied to AI infrastructure still has a late-cycle tailwind. That matters because it concentrates marginal growth in a narrow set of supply chains — chips, servers, power gear, and assembly — which can keep certain exporters strong even if broader Chinese activity stalls. The bigger second-order effect is that policy easing may not transmit cleanly into equity upside if credit demand stays weak and FX pressure persists.

For the listed names, the most important angle is that any AI-linked China manufacturing beneficiary is increasingly being priced as a quasi-global capex lever, not a China beta trade. That favors suppliers with pricing power and scarce product, while commoditized hardware assemblers remain vulnerable to margin compression if yuan strength and input-cost pressure continue. If Washington-Beijing trade friction cools even modestly, the market will likely re-rate the “AI export” complex faster than it will improve the domestic China basket.

The main risk to the bearish China macro read is a policy gap-filling response over the next 1-3 months: additional liquidity, targeted consumption support, or FX stabilization could stop the downgrade cycle before it turns into earnings cuts. But the contrarian point is that a weaker PMI at the threshold does not automatically imply outright contraction in equities; it can actually be bullish for select exporters if domestic stimulus sustains input demand while global AI capex remains intact. In other words, the trade is not “short China,” it is “long the narrow winners, short the broad cyclicals that need a clean domestic rebound.”

Key Decisions for Investors

• Long NVDA vs. short a China domestic-demand proxy basket for 1-3 months: use NVDA as the cleaner expression of AI capex durability while avoiding the weaker transmission from Chinese macro data to broad equities; risk is any export-control escalation or abrupt AI capex pause.
• Long SMCI on pullbacks, but with tight risk controls over the next 2-6 weeks: this is the highest beta beneficiary if AI hardware demand stays strong, yet it is also most exposed to margin volatility and supply-chain hiccups; consider call spreads rather than outright equity.
• Avoid adding to GS/NMR solely on China stimulus hopes for now: the macro backdrop is not yet strong enough to justify a cyclical reacceleration trade, and any benefit from policy support is likely to show up with a 2-4 month lag rather than immediately.
• Pair trade: long NVDA / short a basket of China-sensitive industrials or broad EM cyclicals for 1-2 quarters: the thesis is that AI-linked export demand remains the cleaner earnings driver than marginal domestic China growth, with a favorable risk/reward if PMI stays near stall speed.
• If you want to play a reversal, wait for confirmation: only add risk to broad China beta after a second month of improving non-manufacturing data or a clear RMB stabilization signal; until then, stimulus is more likely to cap downside than create a durable upside trend.