The rise of generative and agentic AI is forcing organizations to rethink their infrastructure. The rising cost of public cloud means that repatriation is happening. The data center never went away, but it is becoming a critical concept for an increasing number of organizations.  

Below are seven themes and talking points concerning the data center in the age of AI. We discussed these issues in the latest episode of the Global Tech Tales podcast, which you can view here or in the YouTube player below: 

What even is the ‘data center’ in 2025?

Without wishing to be too philosophical, in the modern organization the ‘’ is a state of mind. The term is almost synonymous with infrastructure: where your data is, how it’s stored, and accessed.  

For most organizations of any size, especially if they have any legacy, it’s likely to be a mix of physical space with servers, public and private cloud. There may be some compute and storage at the edge. In the end it is all data stored on boxes which require power and connectivity: which is something that is forgotten, often.  

Cloud alone is not the answer…

The one thing we knew about cloud is wrong.  

Were always told that centralizing in public cloud would be more efficient and reduce cost and overhead. But some combination of the nature of business and capitalism, and the complex needs of the data layer in the age of AI, have really shifted that mindset.  

I don’t speak to many IT leaders who see their future organization putting everything in the public cloud. Some combination of cost, privacy, legislation and the need to be flexible means having a data center of one’s own is important.  

And yes, AI is driving a part of this. The need to have next gen hardware dealing with very specific use cases means it’s unlikely cloud provision will be cost effective.  

It’s not just about cost and AI. 

Data sovereignty and legislation in general play a part. There is a lot of risk to consider, and it may be easier and safer to manage that in house. Your cloud provider is likely as secure as you can be, but you don’t control that.  

Then there is sustainability – a subject that seems to have disappeared from the public agenda in many organizations but is most definitely not going away as the major cloud providers quietly slip out sustainability reports showing significant increases in power consumption. And geopolitics. The world feels much less stable than it once did, and your own physical data center is, again, totally under your control. 

…but cloud will never die

Which is not to say that era of cloud is in any sense over. Just that it’s a mix for almost all organizations. Cloud is great for scalability and centralized storage of well ordered data. But there are many other use cases. There is a lot to be said for balancing investment in infrastructure with ongoing running costs. 

Data center skills are an issue

It’s a problem. . Lack of skilled workers in data center management is an issue. It’s also an issue that IT leaders tell me they recognize, but business leaders are less sensitive to.  

Edge computing and AI PCs headed for prime time

The use cases aren’t proven out today, but if you want to build quickly scalable, secure, AI compute in geographically discrete locations the AI PC model theoretically works.  

Don’t think of everyone having an AI PC. Think about specialized AI-driven tasks and roles, with clusters of AI PCs in effect being mini data centers built around tasks and functions.   

Chip makers have their say

 Hardware vendors and the makers of processors have skin in every part of this game. They are driving the AI PC charge, but whether it is your data center or a massive public cloud – it’s run on hardware. And as organizations are figuring out the increased demands of their future org, chipmakers are laughing all the way to the bank. However you build your infrastructure at some point you need next-gen technology. 

Balance is required

In the end as with all aspects of AI, organizations need to find a balance. Given the nature of geopolitics, and climate change, can you afford to place all of your data into the hands of a third party? You need to be able to spread risk both geographically and in terms of inside and outside of your org 

And it’s about opportunities. Get your data structure right – not perfect but right – and then blend the best combination of data center, cloud and end user device edge computing.  Download the .