For years, the overarching question in enterprise infrastructure has been: where should workloads live?
The choices have been simple: cloud or on-premises, centralised or distributed. While those choices remain important, the rise of AI has changed the game, with enterprises today also needing to consider where work itself is taking place and whether their infrastructure is ready to support it there.
The direction of travel has traditionally been towards the centre. Applications migrated to the cloud, and data moved with them, helping businesses to manage their growing volumes of data while simplifying complex environments.
However, this was a model built on a way of working in which users and workloads would come to the data. This is being challenged significantly today as employees now operate beyond the traditional office, while AI needs access to all data across the organisation, no matter where or how it was created.
The question is no longer about where data sits, but whether employees can securely access and use it wherever they are working. And so, edge is back on the agenda.
Bringing the edge back into focus
Looking back a few years, the enforced hybridisation of work completely changed how enterprises operate, and its seismic impact remains. Organisations swiftly adapted to this 'new normal', as work now operates across corporate offices, homes, job sites, and countless other environments. This redefinition of the workplace is placing very different demands on traditional architectures, as enterprises' most valuable data is now created and used outside traditional office environments.
Consider an engineer accessing a multi-gigabyte design file from home. Their infrastructure requirements may not be so different from those of a team operating at a remote job site: both need fast, reliable access to trusted enterprise data. Yet in many cases, they are still being served by infrastructure designed around a far more centralised model of work. Large design files, engineering datasets and rich media assets can quickly expose the limitations of conventional remote-access approaches.
With distributed work becoming an ongoing feature of business operations, infrastructure must be adapted to support this, and that's why the edge is coming back into focus. Often misunderstood as simply a physical location - such as a factory floor, retail environment or IoT deployment - the modern enterprise edge extends much further. It exists wherever people, applications and data come together to perform meaningful work.
In the age of AI, this has become even more important because context matters - AI must have access to information about how data has been created and used to make a meaningful difference. The challenges emerge when data must travel significant distances or be repeatedly copied and moved, as in these scenarios retaining that context becomes more difficult.
The edge is therefore far more than another physical place to deploy infrastructure, but instead an extension of the enterprise itself, allowing intelligence to operate closer to the people, work and data it serves.
Building the data foundation for enterprise AI
As data is created and consumed across a growing number of environments, the economics of moving it are beginning to change. Not only is transferring large datasets back to a central location consistently costly and inefficient, but there is now much greater value in being able to work with data within its operational context.
When a new workflow, application or AI use case emerges, creating another copy of the data can appear to be the quickest route to making it accessible, but when those copies start to multiply, keeping them aligned becomes harder, governance gets more complicated, and teams start asking, 'which version can we actually trust?'
Constant data movement has therefore become increasingly difficult to justify, and enterprises are now beginning to think differently about data access. This new approach focuses on making data securely accessible where it is needed. A solid data foundation enables businesses to introduce new capabilities without needing to redesign their data architecture for every use case.
We must acknowledge that not every environment needs to operate in the same way; locations will have different requirements around performance, connectivity and security. The key is ensuring that the underlying approach to data remains consistent, and this starts by understanding what data businesses have, who or what is accessing it, and that the same governance policies apply regardless of location.
In a current case, a global manufacturer holds petabytes of unstructured file data across dozens of plant sites and legacy NAS systems, including engineering documents, quality records, and standard operating procedures. Its early AI initiatives could reach only a fraction of that estate, limiting model accuracy and pushing teams toward ungoverned workarounds. The company is now consolidating onto a single cloud-backed file platform with a governed index that AI applications can query directly, delivering faster, more reliable answers for engineers and shift supervisors, drawn from authoritative content across every site, rather than isolated pockets of data.
Without developing that greater consistency, distributing intelligence risks creating more complexity rather than reducing it.
Why architecture must be everywhere
This is ultimately what an "everywhere" architecture looks like. It doesn't mean putting infrastructure in every possible location, but giving people and AI secure access to the same trusted data, regardless of where the work takes place.
Agentic AI makes this even more important. Agents need access to enterprise data in the same way that employees do, and architectures that rely on continually moving and duplicating that data will create friction for both.
The edge matters because the enterprise itself has changed. Work no longer happens in one place, and neither does intelligence. It's time that the architecture that supports this new way of working reflects that reality.