Samsung to make AMD and Nvidia HBM parts at second Texas fab, report says
THE SO WHAT
HBM base dies for AMD and Nvidia coming out of Samsung’s second Texas fab is a geographic and political hedge — more AI memory capacity on US soil. If your roadmap is compute-constrained, start tracking HBM supply by region, not just GPU counts, when you negotiate cloud and hardware commitments.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechIndustry experts warn that slow US power grid expansion could stifle its AI boom and give China an edge, as China builds power infrastructure faster
Compute roadmaps that ignore transmission constraints are now fiction—US grid buildout is running at roughly one-quarter the needed pace while China leans into power infrastructure as a strategic lever. If you're planning multi-hundred-MW AI campuses in the US, you need a parallel strategy for grid interconnect risk, on-site generation, and jurisdictional diversification.
Deep & Emerging TechAMD projects ~$17B in Q4 2026 revenue, compared to Intel's $15B to $16B, potentially overtaking Intel as the largest x86 processor vendor for the first time
AMD guiding to ~$17B vs Intel’s $15–16B for Q4 is less about bragging rights and more about who gets design wins for AI-adjacent compute over the next cycle. If you’re locking in server or client roadmaps, treat x86 vendor choice as a strategic hedge on ecosystem, not a default to incumbency.
Deep & Emerging TechSingapore Air Becomes Latest Carrier Affected by Data Breach
One software provider breach exposing 200,000+ flight bookings across airlines is a reminder that aviation’s real risk surface is shared vendors, not just individual carriers. If you run any high-regulation operation, treat core SaaS and PSS vendors as critical infrastructure — demand breach playbooks, segmentation details, and audit rights, not just SLAs.
Deep & Emerging TechQuantum Computing on Commercial Scale
When CEOs from QuNorth, IQM, and Nu Quantum are on stage talking infrastructure and demand, quantum is shifting from lab narrative to early commercial segmentation—who buys what, when, and for which workloads. If you have heavy optimization, crypto, or materials workloads, this is the moment to assign someone to track concrete pilots, not white papers.