Nvidia's next-generation Rubin Ultra accelerator will ship with 192GB of HBM per GPU, down from 288GB in standard Rubin and B300, according to SemiAnalysis, which first reported the change. SemiAnalys…
Amazon Web Services plans to deploy 2 million additional NVIDIA GPUs in 2027 and 2028, on top of more than 1 million GPUs previously planned beginning in 2026, according to an announcement at NVIDIA G…
Amazon Web Services and NVIDIA announced plans to deploy 2 million additional NVIDIA GPUs across AWS data centers between 2027 and 2028, expanding their 16-year collaboration to include NVIDIA Vera CP…
Frore Systems claims its LiquidJet coldplate reduces Nvidia Rubin GPU junction temperatures by 6 to 12°C compared with conventional direct-to-chip cooling, yielding a 10 to 25 percent gain in tokens p…
Nvidia's stock surged more than 10% in the week ending August 8, 2026, easing market fears about AI infrastructure spending, according to analysts. The rally followed SpaceX's announcement of plans to…
Analyst estimates from BofA Global Research and Morgan Stanley project an NVIDIA Rubin Ultra rack costing nearly $21 million, with HBM4e memory alone accounting for about $1.5 million per rack. The fi…
Nvidia's next AI server rack, Kyber NVL144, has been delayed more than a year to 2028 due to circuit board manufacturing problems, according to analyst firm SemiAnalysis. Asian suppliers lost up to do…
Nvidia canceled a planned quad-die version of its Rubin Ultra AI accelerator, shifting to a dual-die design due to manufacturing execution concerns. The change moves from highly integrated CoWoS-L pac…
SK hynix shipped samples of its 12-layer HBM4E memory to major customers, entering customer qualification about three weeks after rival Samsung Electronics. The device offers 48GB capacity, 16Gbps per…
SK hynix has shipped samples of its 12-layer HBM4E memory to major customers, entering the customer-qualification stage weeks after rival Samsung Electronics claimed the industry's first such shipment…