Building The AI Factory Datacenter

AI is changing what “good” looks like in the modern datacenter. It is not just a question of buying faster hardware. Power ceilings, rack density, cooling, and day-to-day operations are becoming first-order design constraints, especially as organizations move from pilots to always-on AI workloads.

In this Video Hot Seat, Tim Phillips talks with Dave Strong, director of advisory and professional services for the UK, Ireland, Middle East and Africa at Hewlett Packard Enterprise, about what that shift means for enterprise planning. Strong describes the change as seismic, and argues that datacenters built for a previous era can become a limiting factor if organizations do not adapt.

A key idea in the conversation is the “AI factory.” In Strong’s framing, you put energy and data in, apply compute, and aim to get actionable insights out. That mindset pushes teams to think beyond servers and storage, and consider facilities, platforms and operations together.

Watch the Video Hot Seat below to hear Strong’s take on what needs to change, and what is worth prioritizing first.

You will learn about:

Full stack integration: why Strong believes combining infrastructure, software layers and services can simplify deployment and operations, and help organizations focus on outcomes rather than stitching together components.

Usage and operational efficiency: where the gains might show up, including how platform choices affect staff effort, complexity, and the ability to run AI consistently.

Energy and heat as design inputs: a grounded look at power availability, how heat is managed, and why some organizations are considering more creative approaches to waste heat.

Modular and prefabricated datacenters: how smaller, repeatable builds can reduce delivery timelines and offer more flexibility, including in constrained environments such as the UK.

Networking and the edge: how AI data movement changes the role of the network, and what to consider when data is generated outside the core datacenter.

AIOps and automation: where operations could become more predictive and less manual, particularly as estates grow more distributed.

If you are responsible for infrastructure strategy, capacity planning, or making AI ambitions real inside existing constraints, this session is a useful watch.