Hormuz Resumes, Micron Surges, and Alibaba Extracts Claude

The overnight cycle runs on three tracks that intersect in ways the individual wire stories do not make explicit. Hormuz traffic is moving again after the container-ship strike triggered the evacuation halt, and oil has retreated to pre-war levels, giving the Hong Kong credit desk a brief window to reprice Hormuz-exposed trade-finance exposure before Friday close. A Singapore-based trade-finance lender re-examines letters of credit on Gulf routes today while the reprieve holds, because the IDF's refusal to leave southern Lebanon means the next escalation catalyst is already in position. The second track is silicon. Micron's forecast, projecting chip shortage conditions past 2027, sent equities up globally and affirmed that the AI-hardware cycle has more runway than the bears argued in Q1. An asset manager rebalancing AI-sector exposure this morning has a cleaner data point than anything from the past quarter. The third track carries the longest tail for this readership. Reuters reported that Anthropic has accused Alibaba of illicitly extracting capabilities from Claude, a foundational AI model developed by the US safety lab. A CISO at any institution running Claude API in their stack should review vendor agreements and data-flow architecture today: the allegation reframes model access as an exfiltration surface, not a service relationship. OpenAI's first custom silicon, built with Broadcom, is a footnote today but sets the trajectory. Germany's rail IT outage is a reminder that legacy OT dependency remains the kill switch for critical infrastructure. Watch the Alibaba-Anthropic filing for any injunctive relief motion before the US weekend.