Espresso

Sign in Subscribe

HPC

Floating Data Centers and AI Power Challenges: Energy, GPU Deals and Quantum Secure

Floating Data Centers and AI Power Challenges: Energy, GPU Deals and Quantum Secure

TL;DR * Floating Data Centers Above Clouds Aim to Cut Energy Use * US Power Shortage Threatens 10% of Data Center Demand by 2030 * 30% Power Demand Growth Forecasts for Data Centers * OpenAI's $1.4 Trillion Compute Commitment * Quantum Security Moves Toward Post-Quantum Infrastructure Floating Data Centers: A Viable

Starcloud launches Nvidia H100 GPUs for orbital data‑center prototypes.

Starcloud launches Nvidia H100 GPUs for orbital data‑center prototypes.

TL;DR * Starcloud to Launch Nvidia H100 GPU for Orbital Data‑Center Prototyping * Nvidia’s $100 B Investment Catalyzes OpenAI and Cloud AI GPU Infrastructure * Liquid‑Cooling Costs Surge to $50k+ for High‑Power GPU Trays in Modern Data Centers * OpenAI’s Transition to Vendor GPU Farms Driven by Massive

Cloud GPUs Slash AI Training Costs; Space-Based AI Centers Soar; World‑Wide Sovereign LLM Rollout

Cloud GPUs Slash AI Training Costs; Space-Based AI Centers Soar; World‑Wide Sovereign LLM Rollout

TL;DR * Cloud GPU rentals cut AI training costs by 95% per‑hour pricing * Satellite‑based AI data centers expand with Starcloud‑1, Starlink, and SpaceX, bringing solar‑powered computing to orbit * Sovereign cloud rollout: Microsoft 365 Copilot delivers localized LLM inference across Australia, India, Japan, and the UK by

Azure HBv5: Memory-First HPC Takes the Cloud Stage

Azure HBv5: Memory-First HPC Takes the Cloud Stage

Bandwidth over Cores – The New Performance Rule * HBM3 per node: 400–450 GB (≈9 GB per core) * Sustained memory bandwidth: 6.7–6.9 TB/s (≈7 TB/s peak) * CPU: 352–368 Zen 4 cores @ 4.0 GHz, SMT disabled * Network: dual-rail 800 Gb/s InfiniBand The numbers speak

AWS, OpenAI, NVIDIA Secure $38B GPU Deployment Deal

AWS, OpenAI, NVIDIA Secure $38B GPU Deployment Deal

TL;DR * AWS and OpenAI secure $38 billion agreement, deploying hundreds of thousands of NVIDIA GPUs for cloud‑based AI training. * Multi‑cloud hybrid infrastructures rise, enabling workload portability and reducing single‑provider failure risk across enterprise data centers. * U.S. data centers consume 4 % of national electricity, projections show

Data centers power HPC with gas turbines, solar, and upgraded hardware for AI scaling.

Data centers power HPC with gas turbines, solar, and upgraded hardware for AI scaling.

TL;DR * Data centers adopt gas turbines, renewables, and 600 MW solar to fuel HPC workloads, backed by $9B renewable contracts. * Mini‑DTX motherboards and full‑copper radiators upgrade HPC hardware, boosting compute performance and cooling. * Remote KVM and gas‑turbine‑powered hyper‑scalers enable scalable AI data center operations.

Qualcomm’s AI 100 Ultra Outperforms Nvidia GPUs In Power Efficiency

Qualcomm’s AI 100 Ultra Outperforms Nvidia GPUs In Power Efficiency

TL;DR * Qualcomm’s AI 100 Ultra chip outperforms Nvidia A50 GPUs in inference per watt * Qualcomm equips datacenter AI with PCI‑Express cards interlinking four AI 100 units * Qualcomm increases AI 100 chip SRAM from 126 MB to 144 MB per chip * Qualcomm’s datacenter AI ambitions remain limited