Super Micro Computer Inc. has introduced what it says is the industry's broadest line-up of enterprise servers for Nvidia Corp.’s forthcoming Blackwell architecture, unveiling more than 30 systems configured with the HGX B200, GB200 NVL72 and RTX Pro 6000 accelerators. The portfolio, now available to customers in Europe, is aimed at speeding the deployment of so-called AI factories; the stock rose about 0.7% in pre-market trading after the 11 June announcement. On 12 June the company extended its offering again, launching both liquid- and air-cooled servers based on Advanced Micro Devices Inc.’s new Instinct MI350 graphics processors. The platforms incorporate Micron Technology Inc.’s 36-gigabyte HBM3E memory and are designed to boost performance while containing power and cooling costs for large-scale AI workloads. Supermicro’s moves come as vendors across the data-centre ecosystem step up their own AI infrastructure plans. Cisco Systems Inc. this week outlined an “AI-ready” data-centre strategy that deepens its collaboration with Nvidia, and Schneider Electric SE announced a partnership with the chipmaker to accelerate the roll-out of AI factories. Cloud providers Vultr and DigitalOcean said they will adopt AMD’s latest MI355X accelerators, while Crusoe plans to purchase $400 million of AMD chips for its AI data centres. The wave of announcements highlights intensifying competition among hardware suppliers to secure next-generation GPUs and develop specialised cooling, memory and networking gear as demand for enterprise AI computing continues to climb.