Memory prices from Samsung, SK Hynix, and Micron are the primary driver of Nvidia's 17% server rack price increases; suppliers are investing in capacity but physical constraints limit near-…

DRAM supply constraints from Samsung, SK Hynix, and Micron are driving 17% price increases on Nvidia's latest server racks; Nvidia invests in memory suppliers to expand capacity but physica…

Nvidia's newest server platforms (Vera Rubin 200, Grace Blackwell 300) face cost pressure not just from GPUs but from memory, storage, and networking components, signaling that AI infrastru…

Memory prices from Samsung, SK Hynix, and Micron are the primary driver of Nvidia's 17% price increase on Grace Blackwell and Vera Rubin server racks; suppliers are physically unable to mee…

Nvidia's strategy of developing and open-sourcing its own models (Nemotron) and investing in application-layer companies like Perplexity and Poolside aims to expand the total addressable ma…

By releasing model weights, training data, and fine-tuning environments for free, Nvidia lowers barriers to AI adoption for companies outside the top hyperscalers, betting that broader mode…

Nvidia and Perplexity collaborate multiple times per week on hardware design, with Perplexity sharing anonymized real-world usage data to inform Nvidia's model training and chip development…

Nvidia's upcoming open model (~1 trillion parameters) will be its largest yet but still smaller than frontier models; the key questions are whether Poolside talent can make it competitive w…