Micron could dethrone Nvidia and become the biggest driver of S&P 500 profit growth
THE SO WHAT
If Micron’s memory economics start driving more S&P 500 profit growth than GPUs, the AI trade is broadening from compute to the full memory and storage stack. Infra buyers should revisit their TCO models—HBM, DRAM, and bandwidth constraints may dictate architecture choices as much as accelerator selection.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechGlobalFoundries reportedly sees strong demand for Chinese optical modules
Surging demand for Chinese optical modules as more data centers go up means bandwidth, not just compute, is becoming a chokepoint in the AI buildout. Infra teams should be modeling optical interconnect availability and vendor concentration risk alongside GPU supply.
Deep & Emerging TechCouncil backed Devon AI datacentre before it went public, Guardian reports
Quiet pre-approval of a large AI data center in Devon shows how much of the AI buildout is being negotiated in back rooms long before communities hear about it. If your operations depend on new data center capacity, track local planning processes and FOIA-driven disclosures—permitting risk is now as material as chip supply.
Deep & Emerging TechAoostar Maco 6850H mini PC review
More Ryzen-powered NUC-class boxes with decent performance widen the options for local inference, edge analytics, and lab clusters. If your workloads don’t need hyperscaler glue, small form factor PCs are becoming a viable part of the infra mix.
Deep & Emerging TechWhy the Coldcard and Liquid Network Hacks Were the Worst in Bitcoin’s History
Trusted Bitcoin hardware and network layers getting compromised is a warning shot for anyone treating specialized infra as inherently safe. If you’re building on crypto rails, assume even “battle-tested” components can fail and design for blast radius, not just prevention.