Enterprise AI is moving beyond the data center. The first wave of AI has been mainly about analyzing information, automating knowledge work and improving decisions. The next wave will connect AI to products, operations and the physical world.
An increasing share of AI’s business value will come from combining analysis and the ability to act in real time. AI will not run in just one place. Some workloads will stay in the large data centers, some will move closer to where data is generated, and others will run directly on devices, machines, vehicles and enterprise systems. This will change the infrastructure question for enterprises.
The question becomes how to make intelligence, compute, data and connectivity work together across all these environments. This is where the intelligent fabric comes in.
The intelligent fabric is an open, secure and programmable network infrastructure that enables distributed AI systems to work together in a more intelligent way and with predictable performance. As intelligence becomes embedded throughout the network itself, it allows networks to optimize autonomously and expose capabilities directly to applications. The network is no longer only moving data; it becomes an active part of how AI systems operate.
For enterprises, that is becoming tangible as manufacturers run some AI directly in machines, computer vision at the edge and production optimization in the data center. The intelligent fabric will enable things like massive-scale digital twins, which are accurate virtual versions of machines or entire cities. They will allow businesses to test complex tasks in a live simulation before acting in the real world.
A logistics company may combine connected vehicles, distributed compute and cloud-based AI to continuously optimize operations. A retailer may use AI across stores, warehouses, devices and central platforms while keeping time-critical or sensitive processing closer to where it happens. Physical AI adds another dimension, as robots, drones and autonomous machines gain the ability to sense, decide and act in the physical world.
At Hitachi Astemo’s Kentucky plant, Ericsson, Hitachi and AWS combined private 5G, edge-to-cloud technology and AI-based computer vision to inspect 24 production components simultaneously, compared with one at a time previously.
At Singapore’s National University Health System, 5G already supports holographic surgical planning, remote surgical assistance, intelligent robotics and hospital-at-home services, connecting intelligence and care across locations.
The strategic implication and why 5G matters
Enterprises should not design their AI strategy around the assumption that intelligence will remain centralized. They will have the ability to build open and flexible architectures where AI workloads can run across devices, cloud, enterprise infrastructure and distributed compute. Networks will provide the connectivity, programmability and rich capabilities needed to bring those environments together.
In enterprises, 5G’s role is different than that of existing fiber and Wi-Fi. 5G extends high-performance and programmable connectivity into environments where intelligence is distributed and increasingly embedded in business operations.
5G provides capabilities such as distributed intelligence with predictable performance, enhanced security, and of course mobility. But the shift goes beyond connectivity. At the core of the intelligent fabric, the mobile network is evolving into a platform with an expanding set of capabilities that applications can use.
Today this includes fraud prevention, differentiated quality of service, high-capacity uplink and positioning. Over time, distributed compute, increasingly precise spatial information, sensing and other capabilities will become part of that platform. Through network APIs, these capabilities are exposed directly to developers and agents for applications to leverage.
This changes the role of the network in the application architecture. For example, one critical application may require guaranteed performance. Another may prioritize cost. Data may need to remain local for security, governance or resilience reasons. Or a workload may need nearby compute rather than a distant cloud.
The intelligent fabric gives enterprises the flexibility to support these choices – allowing the network to adapt and expose the capabilities applications need.
From connectivity to business capability
This is why 5G cannot be treated simply as an IT or network procurement decision. As intelligence becomes more deeply embedded in products and operations, connectivity is at the core of what enterprises can design, build and sell.
It can affect how flexibly a factory can be reconfigured, how intelligent products evolve after they leave the factory, how employees interact with AI, and how companies create new services around connected products and operations. The strategic question becomes even broader. Can my enterprise place intelligence where it creates the most value, while allowing devices, compute, cloud and networks to work together as one environment?
That is as much a business architecture decision as it is a network one.
Building toward 6G
5G is just the starting point for this evolution. With 6G, the intelligent fabric will deepen the convergence of AI, connectivity and distributed compute. Networks will be AI-native and expose a richer set of capabilities to applications and AI systems. The network will provide enterprises’ applications and business processes with ubiquitous tailored performance at scale.
Integrated sensing and communication will be one important step as networks go beyond connecting devices. The network will be able to sense aspects of the physical environment and provide new forms of spatial information. This will create new possibilities across manufacturing, transport, logistics, public safety and other industries where spatial awareness and real-time interaction matter.
Why this matters now
The architectural choices enterprises make over the next few years will shape how easily they can scale AI later. The decisions that executives make around data architecture, compute placement, connectivity and open platforms can either preserve strategic flexibility or create new constraints.
The objective is not to predict where every future AI workload will run. It should be to create an open architecture that allows those choices to evolve as technology, regulation and business models change.
Enterprises and societies need an intelligent fabric that can connect, adapt, sense and expose the capabilities AI applications need, wherever the business needs them.
Read more:
