{"id":326056,"date":"2025-10-23T08:47:14","date_gmt":"2025-10-23T08:47:14","guid":{"rendered":"https:\/\/www.europesays.com\/us\/326056\/"},"modified":"2025-10-23T08:47:14","modified_gmt":"2025-10-23T08:47:14","slug":"dcai-expands-ai-infrastructure-offering-with-wekas-integrated-storage-services-corporate","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/us\/326056\/","title":{"rendered":"DCAI Expands AI Infrastructure Offering with WEKA’s Integrated Storage Services | Corporate"},"content":{"rendered":"

WEKA<\/p>\n

\/ Key word(s): Product Launch<\/p>\n

DCAI Expands AI Infrastructure Offering with WEKA\u2019s Integrated Storage Services <\/p>\n

23.10.2025 \/ 10:05 CET\/CEST
The issuer is solely responsible for the content of this announcement.<\/p>\n

\n DCAI Delivers Hyperscaler-Grade Performance Across Storage and Compute, Complementing Denmark\u2019s Gefion Supercomputer with WEKA\u2019s NeuralMesh Technology\n <\/p>\n

\n COPENHAGEN, Denmark and CAMPBELL, Calif., Oct. 23, 2025 \/PRNewswire\/ \u2014 The Danish Centre for AI Innovation (DCAI<\/a>), operator of Gefion\u2014one of Europe\u2019s most powerful AI supercomputers\u2014and WEKA<\/a>, the AI storage company, today announced the launch of one of Europe\u2019s first fully sovereign, end-to-end encrypted AI infrastructure that unifies high-performance compute and storage in a single platform. The deployment marks a watershed moment for European AI innovation, delivering hyperscaler-grade capabilities with guaranteed data sovereignty.<\/p>\n

\"DCAI<\/p>\n

With this additional 140 petabytes (PB) of usable storage capacity\u2014including 80 PB of performance-optimized NVMe SSDs running on WEKA\u2019s NeuralMesh\u2122<\/a> storage software\u2014Gefion now provides a complete AI factory where European organizations can train models, manage data, and deploy intelligent systems without compromising on sovereignty, security, or performance.<\/p>\n

\n Unifying Compute and Storage
<\/b>Until now, customers relied on DCAI primarily for access to cutting-edge compute power to train and deploy their AI models at scale. With the introduction of this storage service, they can now\u00a0manage their entire AI workflow \u2014 from data ingestion to model deployment \u2014 within a single trusted infrastructure.<\/p>\n

\u201cOur mission is to give innovators the infrastructure they need to build the future,\u201d said Nadia Carlsten,\u00a0CEO at DCAI. \u201cBy adding storage to our platform, we\u2019re making it easier for customers to run end-to-end AI workloads in one place \u2014 with the same level of performance, sovereignty, and security they\u2019ve come to trust from our compute services.\u201d<\/p>\n

WEKA\u2019s NeuralMesh architecture delivers unprecedented performance metrics for AI infrastructure, delivering sub-millisecond latency for AI training and inference workloads while eliminating traditional data input\/output (IO) bottlenecks that impact GPU utilization and performance.<\/p>\n

\n Redefining AI Infrastructure Economics and Efficiency
<\/b>By pairing WEKA\u2019s NeuralMesh software with Gefion\u2019s NVIDIA GPU cluster, DCAI has created a zero-copy, high-throughput data path that continuously feeds GPUs with data. This architectural approach transforms effective GPU utilization by enabling true sovereign multi-tenant orchestration, ensuring compute capacity is continuously leveraged across research, enterprise, and government workloads without compromising security or data isolation.<\/p>\n

The integration also delivers substantial cost and sustainability benefits. DCAI\u2019s single-tier, software-defined architecture reduces infrastructure footprint, energy consumption, and operational complexity compared to traditional multi-tiered systems. By eliminating the need for data replication or external cloud egress, customers can significantly reduce their total cost of ownership while advancing their green AI objectives.<\/p>\n

\n Accelerating AI Innovation at Scale
<\/b>The platform also plans to leverage WEKA\u2019s Augmented Memory Grid\u2122 technology in the future to deliver exceptional time-to-first-token performance for large language models and inference workloads, enabling customers to:<\/p>\n