SEMI CEO on Memory Prices, Semiconductor Supercycle
THE SO WHAT
SEMI’s CEO talking about a memory supercycle underscores that AI demand is now a structural driver of DRAM and NAND pricing, not a side case. If your unit economics depend on storage or memory-heavy workloads, lock in terms where you can and model 2027–2028 as tight, not cheap.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechA botnet running for 23 years with over 15,000 endpoints has finally been shut down by law enforcement and Crowdstrike
A 23-year-old botnet with 15,000+ endpoints staying useful until 2026 is a reminder that legacy compromises don’t just fade out. If you’re rolling out AI or cloud workloads, treat inherited network and endpoint debt as an active risk surface, not background noise.
Deep & Emerging TechBetting on AI and Robots to Automate Superconductor Discovery
AI plus robotics targeting superconductor discovery is about compressing decades of materials R&D into tighter feedback loops. If you’re in any materials-heavy sector, expect the bar to move from “we run simulations” to “we run closed-loop, automated experimentation” over the next few planning cycles.
Deep & Emerging TechThermal Master P3 review: Excellent close-up thermal imaging for your phone
Phone-pluggable thermal cameras with sub-35mK sensitivity and millimeter-range focus turn any field worker into a basic inspector. If you run distributed operations — facilities, utilities, logistics — this kind of hardware is now cheap enough to standardize into SOPs instead of reserving for specialists.
Deep & Emerging TechPsiQuantum and Brookhaven Lab Partner on Fault-Tolerant Quantum Algorithms
PsiQuantum handing its Construct platform to Brookhaven to explore fault-tolerant algorithms is about seeding the future workload pipeline before hardware fully arrives. If you’re in HPC-heavy science or finance, it’s time to at least map which of your problems might migrate to early quantum once it’s practical.