Why is the US more exuberant than China?

U.S. industrial production rose just 0.1% in May, but technology-related manufacturing, semiconductors, and AI infrastructure continue to support the U.S. outlook. China remains weak, with April real retail sales down 1.2% year over year, industrial production up 4.1%, and bank loan growth slowing to 5.5% in May, the weakest pace since 2001. The article suggests a persistent divergence between U.S. strength and Chinese domestic-demand softness, with AI data-center power demand and higher defense spending as longer-term positives.

Analysis

The key market implication is not simply "U.S. good / China bad," but that global industrial excess is likely to stay disinflationary while U.S. capex remains selectively inflationary. That mix favors firms with pricing power tied to scarce compute, grid capacity, and defense procurement, while punishing upstream industrials and commodity-linked exporters exposed to weak Chinese end demand and forced export competition. In other words, the margin pressure is likely to migrate from China-facing manufacturers into the rest of the world through lower export prices and more aggressive discounting.

The second-order winner is the power stack. AI data-center load growth is a multi-year demand call on utilities, gas turbines, switchgear, cooling, and transmission equipment, but the bottleneck is execution, not demand: interconnection queues, transformer lead times, and permitting can turn a secular tailwind into delayed earnings recognition. That suggests the best positioning is in the "picks and shovels" around grid buildout rather than the broad utility basket, because regulated utilities often lag the capex cycle while equipment vendors can capture near-term order backlogs.

For China, the weak consumer/strong factory combination raises the probability of an export-led margin reset in Asia rather than an immediate policy-driven domestic rebound. The risk is that any stimulus lands too late to rescue household balance sheets, which means cyclicals tied to internal demand may stay trapped for months even if headline GDP stabilizes. A more nuanced bullish angle is on firms that benefit from Chinese overcapacity elsewhere: U.S. and European buyers of cheap industrial inputs, and select retailers with inventory flexibility, could see gross margin relief if the deflationary export wave persists.

The contrarian view is that markets may be underestimating how much of the AI and defense capex story is already crowded. If power demand and industrial production do not accelerate further over the next 1-2 quarters, the revenue re-acceleration thesis for many adjacent names could de-rate quickly. The cleaner asymmetry is not in owning "AI" broadly, but in owning constrained infrastructure capacity and avoiding companies whose order books depend on a sustained reopening in China that has yet to materialize.

Key Decisions for Investors

• Long VST / NRG on a 6-12 month horizon: benefit from rising incremental power demand tied to AI and data-center load. Use pullbacks to add; risk/reward is favorable if power prices stay firm and capacity tightness persists.
• Long PWR or ETN vs short a China-exposed industrial ETF over 3-6 months: pair the U.S. grid-build beneficiaries against firms more exposed to global manufacturing oversupply. Target 2:1 upside if AI capex continues while Chinese demand stays soft.
• Initiate a long LRCX / short FXI or MCHI pair for 3-9 months: semicap equipment should outperform if U.S. tech capex remains resilient, while China broad equities remain pressured by weak domestic demand and overcapacity. Tight stop if Beijing launches credible household-support stimulus.
• Buy call spreads on SMH into the next 1-2 quarters: the near-term catalyst is continued strength in high-tech manufacturing and AI infrastructure spend, but cap position size because the market is already leaning bullish on the theme.
• Avoid or underweight broad China consumer cyclicals and commodity-sensitive EM exporters until there is evidence of domestic credit reacceleration; the current setup still favors deflation export spillover over a clean internal-demand inflection.