Amkor is taking its Arizona chip packaging campus to $12bn
THE SO WHAT
Advanced packaging is becoming a US-based strategic asset—Amkor’s ~$12B Arizona build-out adds 60,000 m² of cleanroom on top of 33,000 m² already planned, reducing reliance on overseas OSAT capacity. If your roadmap depends on high-bandwidth chiplets or 2.5D/3D stacks, start mapping packaging constraints and domestic options alongside wafer supply.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechCyber criminals keep extorting US law firms
Ongoing extortion of major U.S. law firms—complete with publication of private documents—confirms that professional services are high-value, soft targets. Any firm handling sensitive client data should treat ransomware and data theft as inevitabilities and invest in segmentation, offline backups, and clear breach playbooks, not just perimeter tools.
Deep & Emerging TechMicrosoft’s September updates fix a record 973 security flaws
A record 973 vulnerabilities patched in one month means your Windows estate is effectively a rolling zero-day surface until your patch pipeline catches up. Security and IT ops need tighter coordination—automated deployment, staged rollouts, and clear exception handling—because manual, ticket-driven patching cannot keep pace with this volume.
Deep & Emerging TechYour favorite app is silently collecting your browsing history, Surfshark warns
Surfshark’s finding that 45% of top mobile apps collect browsing history—especially social platforms—raises the bar on what “private” really means in your data map. If your product embeds third-party SDKs or ad tech, assume regulators and users will start treating this kind of collection as surveillance and audit your dependencies before they audit you.
Deep & Emerging TechKawasaki signed an AI chip deal with a startup the Pentagon backs
EdgeCortix landing Kawasaki’s aerial defense platforms under a 2026–2028 teaming agreement shows defense-grade AI compute is fragmenting beyond the usual US hyperscaler and GPU incumbents. If you build for edge or defense, you now have to design and optimize for a more diverse accelerator landscape—start abstracting your inference layer.