The chip industry has a warning for Trump: hands off the memory market
THE SO WHAT
When SEMI tells the White House that intervening in DRAM/NAND could worsen shortages, it’s a reminder that policy risk is now a first-order input to hardware planning. If your roadmap depends on memory pricing or availability, scenario-plan around regulatory shocks—not just foundry capacity and demand cycles.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechUS Energy Dept tells data centers to use backup generators during heatwave to free up power for AC usage
Data centers are now treated as flexible load in grid emergencies—told to burn diesel so households can run AC. If you operate large compute, you need a power strategy that assumes regulatory curtailment and reputational scrutiny, not just a PPA and a backup genset.
Deep & Emerging TechThe NHS plans to offer rewards to people who walk for 30 minutes a day — here's everything we know, including the tech you'll need to take part
State-backed health systems are starting to pay for behavior change, not just treatment—using consumer wearables as the tracking layer. If you’re in health, insurance, or fitness tech, this is a wedge for reimbursement-linked products and new data-sharing rails with public providers.
Deep & Emerging TechEnd of an era — the historic PlayStation factory that manufactured 24 billion gaming discs is already being turned into a micro-optics lab, just days after Sony announced its digital push
Sony repurposing a plant that pressed 24B discs into a micro‑optics lab is a clean example of legacy media capex being recycled into next-gen hardware. If you own physical distribution infrastructure — plants, logistics, retail — start mapping how it can be redeployed into components for sensing, displays, or compute instead of trying to extend declining formats.
Deep & Emerging TechResearchers Propose Thermodynamic Computing Architecture That Could Dramatically Reduce AI Energy Use
A probabilistic, thermodynamic computing architecture built from conventional transistors is a reminder that the escape valve on AI power demand may come from new compute primitives, not just more GPUs. If you’re planning multi‑year AI infra, start tracking non-von Neumann roadmaps alongside your GPU strategy — they could change the economics mid-horizon.