Inside the AI Data Center: An Illustrated Guide to the Engineering, Servers, Networks, Power, and Cooling Behind Modern AI - Tapa blanda

Sinopsis

What does it take to turn an AI workload into a working data center?

Behind every trained model and generated response is a physical system of accelerators, servers, networks, electrical equipment, and cooling infrastructure. Understanding the individual parts is only the beginning. The real engineering challenge is understanding how they work together.

Inside the AI Data Center — 2026–2027 Edition takes you from the chip to the campus, connecting AI workloads to the infrastructure that makes them possible. Across 40 chapters and 101 illustrations, it combines clear explanations, engineering drawings, worked calculations, and practical design examples.

Follow an AI workload through the building: software, chips, networks, electrical systems, and cooling.

Explore what’s inside:

• AI workloads and infrastructure requirements: How training, inference, model size, precision, and performance targets shape facility decisions.
• Chips, servers, and racks: GPUs, specialized accelerators, memory bandwidth, server architecture, rack power, cabling, and service access.
• Networks and storage: Scale-up connections, cluster fabrics, Ethernet and InfiniBand, network topologies, optics, checkpoints, and data pipelines.
• Power and electrical engineering: The utility-to-chip power chain, capacity planning, UPS systems, generators, rack distribution, and electrical safety principles.
• Cooling and heat rejection: Airflow management, direct-to-chip liquid cooling, coolant distribution units, immersion cooling, cooling towers, dry coolers, and heat reuse.
• Site selection and construction: Power availability, connectivity, building layout, structural constraints, engineering coordination, commissioning, and handover.
• Software, operations, and security: Workload scheduling, monitoring, reliability, maintenance, incident response, and physical and cybersecurity boundaries.
• Economics and sustainability: Capital and operating costs, utilization, energy efficiency, water use, emissions, and planning for future hardware.
• Worked design studies: An AI training cluster, a distributed inference facility, an existing-facility retrofit, and a complete system design review.
• Practical reference appendices: Glossary, formulas, unit conversions, sample calculations, diagram-reading guidance, checklists, equipment schedules, and runbooks.

See the equipment. Follow the connections. Understand the numbers.

Technical sidebars explain the arithmetic behind decisions—from accelerator memory and network bandwidth to electrical demand, coolant flow, and cost per useful hour. Cutaways, exploded views, schematics, and system diagrams make complex infrastructure easier to visualize.

Written for engineers, technical managers, developers, students, and infrastructure professionals, this book helps readers build a shared understanding across disciplines. You do not need to be an expert in every field to follow the journey.

Whether you are entering AI infrastructure, evaluating a project, or expanding your engineering knowledge, Inside the AI Data Center gives you a practical foundation for understanding how workload requirements become physical systems—and how to check whether those systems work together.

"Sinopsis" puede pertenecer a otra edición de este libro.