AI infrastructure every day replace — information from ASUS, Block and extra

Block will deploy the NVIDIA DGX SuperPOD and Cerebras expands AI inference information facilities

On this common replace, RCR Wi-fi Information highlights the highest information and developments impacting the booming AI infrastructure sector

ASUS showcases AI options at CloudFest 2025

ASUS unveiled a complete lineup of AI infrastructure options at CloudFest 2025, integrating Intel Xeon 6 processors, NVIDIA GPUs and AMD EPYC chips. The corporate additionally launched RS700-E12, RS720Q-E12 and ESC8000-E12P-series servers, optimized for scalable AI coaching and inference. ASUS additionally debuted the Intel Gaudi 3 AI accelerator PCIe card, designed for environment friendly generative AI inferencing.

Block deploys NVIDIA DGX SuperPOD for open-source AI

Block stated will probably be the primary North American firm to deploy the NVIDIA DGX SuperPOD with DGX GB200 programs, hosted at an Equinix AI-ready information middle. This high-performance infrastructure might be devoted to open-source AI mannequin analysis and coaching, specializing in generative AI improvements in underexplored fields. Block’s AI analysis staff stated it goals to push AI boundaries whereas sustaining a dedication to open-source improvement. The deployment will leverage Lambda 1-Click on Clusters, providing speedy entry to interconnected NVIDIA GPUs for environment friendly large-scale AI experimentation and innovation, the agency added.

Cerebras expands AI inference information facilities

Cerebras Methods introduced the launch of six new AI inference information facilities throughout North America and Europe, outfitted with hundreds of Cerebras CS-3 programs. These amenities will ship over 40 million Llama 70B tokens per second, making Cerebras the most important supplier of high-speed AI inference globally, based on the agency. New areas embrace Minneapolis, Oklahoma Metropolis and Montreal, with extra websites within the Midwest, East Coast and Europe slated for the final quarter of 2025. The Oklahoma Metropolis facility will characteristic 300+ CS-3 programs in a Stage 3+ information middle, optimized with customized water-cooling options for high-efficiency AI processing, guaranteeing world entry to sovereign AI infrastructure, the agency added.

What’s a GPU cluster?

In one other article, RCR Wi-fi Information defines a GPU cluster and its position in AI infrastructure. A GPU cluster consists of interconnected computing nodes, every outfitted with GPUs, CPUs, reminiscence and storage. These nodes talk through high-speed networking, enabling environment friendly information distribution and processing. GPU clusters are on the core of contemporary AI infrastructure, offering the computational energy needed for deep studying, NLP and superior AI-driven functions. As AI continues to redefine industries, organizations that put money into scalable, environment friendly GPU clusters might be well-positioned to capitalize on the complete potential of synthetic intelligence.

Why these bulletins matter?

These developments spotlight the speedy enlargement of AI infrastructure, pushed by rising enterprise demand for scalable, high-performance AI options. With growing AI complexity, corporations are investing in specialised AI {hardware}, liquid-cooled information facilities and hyperscale computing to speed up AI analysis, optimize mannequin coaching and drive next-generation innovation.