Semiconductor sell-off is driven by leveraged ETF unwinding and sentiment, masking robust AI capex fundamentals like TSMC's $100B expansion and surging DRAM spot prices. Chinese AI models spark debate between hardware efficiency risks and Jevons paradox compute demand. US-Iran escalation pushes oil above $90.
The 30-40% correction in memory and semiconductor stocks is heavily attributed to forced liquidations of single-stock leveraged ETFs and CTA threshold breaks, rather than fundamental deterioration. TSMC's $100B Arizona expansion and upward revisions in hyperscaler capex by major banks contradict the peak-AI narrative.
Moonshot AI's Kimi K3 and Alibaba's Qwen are matching US frontier models at lower costs. Bears argue this software efficiency destroys premium hardware demand. Bulls counter with the Jevons paradox: cheaper inference drives exponential usage, ultimately requiring more total compute and forcing US hyperscalers to accelerate spending.
Escalating strikes and threats to the Strait of Hormuz have pushed Brent crude above $90. This geopolitical risk premium is tightening global supply, benefiting US energy producers and PADD 3 refiners, while simultaneously threatening the broader market with renewed inflation and higher-for-longer Fed policy.
Traders are accumulating PADD 3 refiners and US energy as a geopolitical hedge against Strait of Hormuz disruptions and accelerating crack spreads.
Institutional buyers and retail are accumulating neo-cloud and optical components, betting that inference demand will outstrip capacity.
Waiting for forced deleveraging of single-stock ETFs to complete. Foreign futures accumulation suggests a floor, but ADR conversion risks remain.
Waiting for specific technical entry points (Gold below $3,500, Silver below $50) as current geopolitical spikes are weighed against rate-hike fears.
Asset managers advise underweighting EM due to over-concentration in the AI rally and a lack of diversification benefits.
Rotation out of software-layer AI into physical hardware and cyclicals due to pricing pressure from open-weight Chinese models.