Stories

  1. Image: AppleInsider

    Apple reportedly close to long-term DRAM supply deal with Samsung

    Apple is close to a long-term DRAM supply agreement with Samsung, Korea's Maeil Business Newspaper reported. The deal would run four to five years from the first half of 2027, into 2031 or 2032, with a minimum supply volume and prices set later based on market conditions. The terms are described as favorable to Samsung, a departure from Apple's usual short-term contracts, as a memory shortage driven by AI data center demand raises component costs. Apple recently signed a similar deal with SK hynix and earlier in 2026 agreed to pay Samsung double for DRAM.

  2. Image: TechPowerUp

    Judge denies Samsung, SK hynix and Micron motion to dismiss DRAM supply conspiracy suit

    U.S. District Judge Noël Wise denied a joint motion by Samsung, SK hynix and Micron to dismiss a class action accusing them of conspiring to restrict conventional DRAM supply. The suit, filed by 17 plaintiffs, alleges coordinated production cuts from late 2022 and a shift of capacity toward HBM; the price rise is put at 697% over 18 months in one account and about 500% in another. The judge rejected the argument that the case reruns an earlier suit dismissed by the Ninth Circuit in 2022, citing newer alleged conduct. The ruling is at the pleading stage and proves nothing yet.

  3. Image: Tom's Hardware

    Sam Altman argues scarce DRAM is better utilized in AI data centers than home PCs

    OpenAI CEO Sam Altman argued that personal computers use scarce DRAM inefficiently compared with cloud data centers. Speaking on the Cheeky Pint podcast, he said many people assume open-source models will let them run AI locally, reducing the need for data centers. Altman said home memory may sit idle much of the time, while cloud systems can maximize utilization amid supply constraints. The episode was released on October 29.

  4. Image: The Verge

    Sam Altman says it’s better for RAM to be in a data center than on your desk

    OpenAI CEO Sam Altman argued that concentrating memory in data centers is preferable to keeping it in personal computers, saying home PCs may leave scarce DRAM underused. In an interview on the Cheeky Pint podcast, he criticized the assumption that open-source models and local inference will reduce the need for data-center capacity. Altman said cloud resources should be used as fully as possible amid constrained supply.

Topics

Publishers