Lambda is deploying a major GPU cluster at Prime Data Centers’ LAX01 facility in Vernon, California. The agreement involves 21MW of capacity across the 33MW data center for Nvidia Blackwell AI infrastructure. This comes just weeks after Lambda secured a multi-billion-dollar contract with Microsoft.
Spanish energy company Forestalia is making a massive pivot into data centers with a €12 billion investment in Zaragoza. The ‘Búfalo Project’ aims to build three facilities that could create over 30,000 construction jobs. This comes as Microsoft and Amazon also expand their presence in the region.
APAC data center operator PDG is making its South Korean debut with a massive new facility. The company will lease space in a newly announced 80MW data center being developed by real estate firm ESR and Wide Creek Asset Management in Incheon.
Australian AI cloud company Firmus has raised AU$500 million in equity funding to expand its data center capacity across multiple locations. The funding values the company at AU$6 billion and will support Project Southgate, aiming for 1.6GW capacity by 2028. Initial construction is underway in Tasma
Microsoft’s Azure HBv5 instances have reached general availability featuring AMD’s custom EPYC 9V64H processors with HBM3 memory. The new instances offer up to 368 vCPUs and deliver staggering 6.7 TB/s memory bandwidth for HPC applications. Performance benchmarks show dramatic evolution from 2019’s
Nvidia has updated its GeForce NOW cloud gaming service with 12 new titles, including major day-one releases like Call of Duty: Black Ops 7 and Anno 117: Pax Romana. The company also expanded its RTX 5080 server upgrades to Phoenix, with Stockholm next in line for the performance boost.
T-Mobile has activated a major solar installation at its largest Czech data center. The project covers rooftops and surrounding areas of the DC7 campus in Prague. It’s part of the company’s broader energy efficiency strategy for its flagship facility.
Internal Microsoft financials reveal OpenAI’s massive inference spending on Azure cloud services. The documents suggest the AI company’s revenue numbers may be significantly lower than previously reported, raising questions about its financial sustainability.
Data centers now distribute power like utilities but often use outdated metal-clad systems. These centralized designs create single points of failure and limit scalability. The solution? Adopt utility-style distributed switchgear.
Nebius has deployed its first Nvidia Blackwell Ultra AI infrastructure in the UK near London. The initial phase includes 4,000 Nvidia HGX B300 GPUs across 126 racks, with another 3,000 coming in phase two. The company aims to reach 100MW of capacity next year.
