Storage squeeze and energy backlash
A Solidigm exec warned that exploding AI demand could trigger storage‑chip shortages through 2030, creating bottlenecks for data‑intensive systems. At the same time, a new poll shows many Americans blame AI data centers for rising electricity costs—an emerging reputational and policy headwind for hyperscale infra. (telecom.economictimes.indiatimes.com) (rrdailyherald.com)
Greg Matson, Solidigm’s senior vice president, told reporters at Nvidia’s developer conference that AI systems coming later this year could require about 35% more storage than previous systems and warned supplies would be tight through 2030, saying “I could sell twice as much as I am today.” (reuters.com) The Reuters report noted those comments followed remarks by SK Group chairman Chey Tae-won that high-bandwidth memory (HBM) shortages tied to AI demand could persist to 2030, signaling pressure not just on NAND but across adjacent memory tiers used in GPU servers. (reuters.com) Matson said Solidigm plans to introduce higher‑density silicon drives later this year and to expand manufacturing output, but added the company doubts it can fully keep pace with the surge in demand for flash storage. (reuters.com) The Center Square Voters’ Voice Poll, fielded by Noble Predictive Insights from March 2–5, 2026, queried roughly 2,659 registered voters nationwide as part of its March tracking release. (yahoo.com) That poll showed 15% of respondents directly blamed AI data centers for rising electricity costs and 22% said companies were using AI data centers as an excuse to raise profits, a combined 37% attributing higher bills to data centers compared with 27% who blamed general inflation. (yahoo.com) The Center Square summary and regional reprints reported partisan and age splits in the data — combined blame rates around 33% for Republicans, 41% for Democrats and 37% for independents, with younger voters more likely to single out data centers. (dailyfly.com) Industry and policy coverage ties the two threads together: analysts and regulators have pointed to AI-driven data‑center expansion as a contributor to rising rates on grids such as PJM, and utilities and state officials in Virginia and New Jersey have publicly debated how to allocate the costs of new data‑center load. (cnbc.com)
