Conference Tracks
Explore the Technologies Powering the AI Infrastructure Ecosystem
From silicon and systems architecture to networking, power, cooling, enterprise infrastructure, and operations, Digital Infrastructure World delivers focused education designed around the technologies and strategies enabling AI at scale.
Education Designed Around the AI Infrastructure Stack
AI is transforming every layer of digital infrastructure. As organizations deploy increasingly complex AI environments, success depends on collaboration across compute, networking, power, cooling, facilities, software, and operations.
Digital Infrastructure World brings together the leaders building this ecosystem to share practical insights, technical expertise, real-world deployments, and forward-looking strategies that help organizations design, deploy, and scale AI infrastructure.
Agenda at a glance: Designed as one.
Four co-designed tracks under one keynote spine, from the chip to the grid.
Every layer influences the next. Our conference tracks explore the complete AI infrastructure stack.
The Rack
Design the 1MW rack, where silicon, power and cooling are engineered as one.
The Rack
The Rack explores where the machine is built: silicon, power and cooling engineered as one, from the 1MW rack to the busbars, cold plates and CDUs inside it, and the ODMs who integrate it.
What you’ll explore
- Cooling chips past 1000W, beyond what air can carry
- Delivering a megawatt to one rack without losing it to heat
- Adding liquid cooling to halls built for air
Company types
Hyperscalers · Neocloud giants · Enterprise & HPC operators · DC operators · The rack supply chain (OEMs, ODMs, integrators)
Job titles
VP Infrastructure · Head of Rack Engineering · Director, Thermal · Head of Power Delivery · Hardware & Platform Architects · Mechanical Design Leads
The Cluster
Run open clusters for AI, where a million GPUs act as one and drive your tokenomics.
The Cluster
The Cluster explores where thousands of racks become one computer: the fabric, optics, software and storage that bind a million GPUs together, and the economics that decide whether it pays.
What you’ll explore
- Keeping GPUs fed: the gap between 50% and 95% utilisation
- Scaling up, out and across sites without the fabric choking
- Choosing optics and interconnect that don’t strand performance
Company types
Hyperscalers · Neocloud giants · AI labs · DC operators · Enterprise & HPC operators · The networking & optics supply chain
Job titles
VP, AI Infrastructure · Network Architects · Head of Cluster Engineering · ML Platform Leads · Principal Engineers · Head of Reliability
The Data Center
Deploy open data centers for AI, where fungible facilities outlast each chip generation.
The Data Center
The Data Center explores the facility built around the machine: how sites are found, designed, built and run, from powered land and build-to-suit to the power chain, cooling plant and controls that hold a gigawatt together.
What you’ll explore
- Securing sites with power, land and water before the window shuts
- Designing facilities that outlast each chip generation
- Building to suit at gigawatt scale: prefab, modular, faster
Company types
Hyperscalers · DC operators · Build-to-suit developers · Colocation & wholesale · Neocloud giants · The facility supply chain (EPC, power, cooling)
Job titles
VP Data Center Construction · Head of Design & Engineering · Head of Critical Facilities · Mechanical & Cooling Plant Engineers · Facility
The Energy
Power the gigawatts, where energy is the strategy, secured ahead of silicon.
The Energy
The Energy explores where the data center meets the grid: generation, connection, storage and the markets behind them, as campuses cross into gigawatts and power decides what gets built and where.
What you’ll explore
- Securing power years ahead: speed to power beats speed to build
- Connecting to a grid that can’t keep pace
- Generating on-site when the grid won’t: gas, microgrids, nuclear
Company types
Hyperscalers · DC operators · Neocloud giants · Utilities & grid operators · IPPs & power developers · The power supply chain (generation, storage, grid)
Job titles
VP Energy & Power · Director, Energy Procurement · Head of Grid Connection · Power Systems Engineers · Head of On-site Generation · PPA & Energy Markets Leads
Five Disciplines from the chip to the grid
The rapid adoption of AI is reshaping how digital infrastructure is designed, deployed, and operated. Organizations must now navigate increasingly interconnected challenges spanning compute architecture, networking, energy, cooling, facilities, software, and operations.
Digital Infrastructure World is designed to help industry professionals understand these interdependencies through practitioner-led education, technical case studies, and peer-to-peer discussions that reflect the realities of deploying AI infrastructure at scale.
Compute
Compute
Compute topics explore the IT that lives in the rack and the cluster: accelerators and AI servers, the 1MW rack as a system, memory, and the software, storage and economics that turn them into a working machine.
What you’ll explore
- Integrating the 1MW rack as one system: compute, memory, power and cooling
- Feeding accelerators with enough memory and HBM bandwidth
- Running clusters reliably when one node can stall the job
Securing the platform end to end, root of trust and - confidential compute
Making inference pay, cost per token, not cost per GPU
Company types
Hyperscalers · Neocloud giants · AI labs · DC operators · Enterprise & HPC operators · The compute supply chain (silicon, servers, memory)
Job titles
VP, AI Infrastructure · Hardware & Platform Architects · Head of Cluster Engineering · ML
Network
Network
Network topics explore the fabric that binds racks and sites into one machine: scale-up, scale-out and scale-across, the optics and switches beneath it, and the software that runs it.
What you’ll explore
- Keeping GPUs fed, the difference between 50% and 95% utilisation
- Scaling up, out and across sites without the fabric choking
- Choosing optics and interconnect that don’t strand performance
- Connecting multiple sites into one training run
- Holding reliability and congestion in check at a million ports
Company types
Hyperscalers · Neocloud giants · AI labs · DC operators · Enterprise & HPC operators · The networking supply chain (switches, optics, NICs)
Job titles
Network Architects · Head of Fabric Engineering · Optical & Interconnect Leads · Principal Network Engineers · Head of
Physical
Physical
Physical topics explore everything physical that holds the machine: the rack and its frame, the building around it, and how both get designed, built and run.
What you’ll explore
- Standardising racks and connectors across vendors
- Deploying validated reference designs instead of bespoke builds
- Building to suit and modular at gigawatt scale
- Designing for density, weight and floor loading
- Modelling the facility as a digital twin and running it through controls
Company types
Hyperscalers · DC operators · Build-to-suit developers · Colocation & wholesale · Neocloud giants · The facility supply chain (EPC, racks, controls)
Job titles
VP Data Center Construction · Head of Design & Engineering · Head of Critical Facilities · Director, Site Selection · Mechanical Design Leads · Head of DCIM & Controls
Cooling
Cooling
Cooling topics explore taking the heat off the chip and out of the building: from cold plates and CDUs at the rack to chillers, heat reuse, and water systems at the facility.
What you’ll explore
- Cooling chips beyond what air can carry
- Distributing coolant reliably at the rack, CDUs and cold plates
- Adding liquid cooling to halls built for air
- Immersion and two-phase for the densest workloads
- Heat reuse and closed-loop water to hit sustainability targets
Company types
Hyperscalers · DC operators · Neocloud giants · Colocation & wholesale · Enterprise & HPC operators · The cooling supply chain (liquid, CDUs, chillers)
Job titles
Director, Thermal · Head of Cooling Engineering · Mechanical & Cooling Plant Engineers · Liquid Cooling Product Leads · Head of Critical Facilities · Head of Sustainability
Power
Power
Power topics explore carrying power from the grid to the chip: high-voltage DC and busbars at the rack, distribution and storage in the building, generation and grid connection beyond it.
What you’ll explore
- Delivering a megawatt to a single rack
- Shifting to 400/800V DC and disaggregated power
- Storing energy and smoothing the violent load swings of AI training
- Securing power years ahead of the build
- Turning the data center into a grid asset
Company types
Hyperscalers · DC operators · Neocloud giants · Utilities & grid operators · IPPs & power developers · The power supply chain (HVDC, storage, transformers)
Job titles
Head of Power Delivery · VP Energy & Power · Power Systems Engineers · Director, Energy Procurement · Head of Grid Connection · PPA & Energy Markets Leads
Every layer influences the next. Our conference tracks explore the complete AI infrastructure stack.
Continue Exploring the Education Program
Explore the complete Digital Infrastructure World conference experience.