AI spending panic hits chips and Big Tech—while Taiwan probes Nvidia chip smuggling

On July 28, 2026, markets digested a sharp wave of AI-related pessimism that spilled from earnings expectations into chip and mega-cap tech valuations. A report highlighted that South Korea’s Kospi-linked sentiment deteriorated, with the Seoul market falling more than 10% as investors questioned elevated AI spending and the competitive pressure from China. In parallel, coverage ahead of earnings framed Meta as facing a “wall of AI pessimism,” while Business Insider reporting said Amazon is winding down most flagship AI models as part of a strategy overhaul. Reuters also noted that Nasdaq futures slid on AI chip worries ahead of pivotal earnings, and that Google’s earlier report had already triggered a sell-off that is now pressuring Amazon, Meta, and Microsoft. Strategically, the cluster points to a widening gap between AI capex narratives and near-term monetization, with investors increasingly treating AI compute as a cost center rather than a guaranteed revenue engine. That financial skepticism is colliding with a more security-driven supply chain reality: Taiwan prosecutors detained a Nvidia employee in an investigation into alleged semiconductor smuggling to China, underscoring how export controls and enforcement are becoming market-moving. Meanwhile, AMD’s AI data center deal with Core Scientific signals that some suppliers are still locking in demand, but the overall tone suggests a more selective procurement cycle. The power dynamic is clear: US-linked AI platform companies and chipmakers face both capital-market scrutiny and tighter geopolitical constraints, while China’s competitive pressure and illicit supply channels raise the stakes for compliance and enforcement. Economically, the immediate transmission mechanism runs through semiconductors, data centers, and cloud capex expectations. A “chip rout” narrative appears across Reuters coverage, implying downside pressure for AI accelerators and related infrastructure supply chains, with Nasdaq futures reflecting risk-off positioning. The Reuters item on Meta and BlackRock partnering on a $14 billion El Paso data center ties the AI buildout theme to financing and real-asset exposure, but the broader sell-off suggests investors may demand faster payback or clearer unit economics. On the policy and industrial side, Poland’s commitment of over $106M to bid for AI “gigafactories” reinforces that European governments are trying to secure manufacturing capacity, potentially supporting equipment and materials demand even as equity markets reprice AI spending. What to watch next is whether earnings guidance and AI model strategy updates validate the pessimism or force a reversal. Key triggers include Meta’s AI spending commentary, Amazon’s explanation for winding down flagship models, and any additional disclosures from Microsoft and other hyperscalers on compute utilization and margins. On the security front, the Taiwan smuggling investigation—especially any follow-on charges, evidence of diversion routes, or export-control violations—could tighten compliance and disrupt component flows. For markets, watch AI chip pricing signals, data center financing spreads, and guidance on capex intensity into late 2026; escalation risk rises if enforcement actions broaden beyond one case or if additional “chip rout” headlines emerge before earnings season peaks.

Geopolitical Implications

• 01

  Enforcement of export controls is becoming a market variable for AI supply chains.

• 02

  US-linked AI firms face dual pressure from investors and geopolitical compliance constraints.

• 03

  China competition is intertwined with illicit supply-channel risk, raising enforcement stakes.

• 04

  EU industrial policy (AI gigafactories) signals strategic manufacturing competition.

Key Signals

• —Earnings guidance on AI spending and compute utilization from Meta and Amazon.
• —Any expansion of the Taiwan Nvidia smuggling case and export-control findings.
• —AI chip pricing/order visibility and procurement selectivity by hyperscalers.
• —Financing and capex intensity signals tied to large data center projects.