AI Memory & Storage

Raspberry Pi's fifth price round since December 2025 shows the DRAM shortage is still passing straight through to makers of low-cost computing hardware. The 16GB Pi 5 is up 221 percent since launch, and Compute Module 5 variants rise $40. Record first-half revenue of $256.9 million was helped by memory bought before the spike, and that cushion has largely gone. With SK Hynix and Micron warning of shortages into 2028-2030, memory suppliers keep pricing power and hobbyist and embedded buyers face lasting cost pressure. Higher prices may push buyers toward lower-RAM boards, older Pi models, or alternatives, and give grey-market resellers room to exploit shortages.

Fujifilm's fivefold LTO capacity expansion signals confidence that AI training and inference data will drive archival demand for tape. The ¥10 billion outlay is modest, but it matches the LTO Program's forecast of record 2026 shipments after the 2025 dip. It benefits the LTO ecosystem, including HPE, IBM and Quantum, which run the LTO Program and sell drives and libraries. It also reinforces tape's role as the cheapest offline and nearline tier beneath flash and disk in AI data pipelines. Execution risk remains because demand is cyclical and the expansion is sized against forecasts rather than contracted orders.

SK Hynix's warning that the memory shortage will persist through 2030, alongside record profit, signals prolonged pricing power for HBM and DRAM suppliers. Long-term agreements covering more than half of 2027-2028 capacity give the company unusual earnings visibility. Electronics makers such as smartphone, PC and console vendors face continued margin pressure and price hikes. Korean memory shares rallied, with SK Hynix up 10.9%, Samsung up 9% and the KOSPI up 6.6%. The upside is tempered by investor fears of an AI spending slowdown: SK Hynix remains about 50% below its mid-year peak, it postponed the Solidigm IPO, and the Bank of Korea flagged up to $100 billion of export downside.

Applied Materials' $5 billion EPIC center is the largest U.S. investment in semiconductor equipment R&D and puts rival chipmakers beside its engineers with early access to new tools. Faster development cycles, from 10-12 years down to as little as six, would deepen Applied's role in each new chip architecture and lock in its customers. The shares are already up over 110% in 12 months, and the company says it has demand visibility into 2030. Some observers warn that memory-industry boom-bust history and the gap between visibility and signed contracts leave overinvestment risk. CXMT's claimed memory density breakthrough, which sent Samsung and SK Hynix shares down about 7% and 10%, shows the competitive pressure and the export-control stakes.

The deal signals a structural push to localize advanced semiconductor packaging in the US, a step long seen as the weakest link in domestic AI chip manufacturing. GlobalFoundries gains a new high-value revenue stream and strategic relevance in leading-edge packaging alongside TSMC and Intel Foundry. Nvidia, AMD and other HBM-dependent chipmakers stand to benefit from reduced reliance on Taiwan round-trips once capacity comes online in 2028, though near-term bottlenecks persist. The announcement intensifies competitive dynamics among Amkor, Intel Foundry and GlobalFoundries for US CoWoS-compatible packaging business. Investors should view this as a multi-year infrastructure build rather than an immediate capacity relief, given the 2028 timeline.