- Enterprises are increasingly asking operational questions rather than connectivity questions
- AI's next phase will occur in factories, mines, ports, airports and other physical environments
- The unified infrastructure era is dissolving the boundaries between telecom, enterprise IT, cloud, AI and industry
"The biggest shift is, I think, [the network] is not a utility. It becomes part of how operations run."
With that observation, Ericsson's Åsa Tamsons may have captured the single most important change occurring in enterprise technology today.
For decades, enterprise networks were viewed as infrastructure. Important, certainly, but largely invisible to the business itself. Applications created value. Operations generated revenue. Connectivity linked the two together.
That distinction is beginning to disappear.
In a recent conversation with Fierce Network TV, Tamsons, SVP and Head of Enterprise Wireless Solutions at Ericsson, described a market evolving far beyond traditional connectivity. Increasingly, enterprises are no longer asking connectivity questions. They are asking operational questions. They want to understand how to maintain continuity across distributed environments, support automation at scale and ensure that critical systems do not become bottlenecks as operations become more dependent on software, data and machine intelligence.
Those questions are reshaping enterprise networking and, in the process, redefining the role of the network itself.
Modern industrial operations increasingly rely on tightly integrated combinations of automation, robotics, machine intelligence, digital twins, autonomous systems and real-time decision-making platforms. Whether in a mine, a port, an airport, a warehouse or a manufacturing facility, operational performance is becoming inseparable from digital performance.
In those environments, network performance becomes an operational metric.
"The operation needs to be running all the time," Tamsons explained. "If connectivity fails, operations stop, production stops."
That reality changes the nature of the conversation.
The value is no longer found primarily in the connection itself. It increasingly moves up the stack, residing in the ability to orchestrate operations, assure performance, maintain resilience and deliver the operational continuity upon which modern enterprises depend.
Success is measured not by whether devices can communicate, but by whether operations continue to run.
This represents a fundamental shift in how enterprises think about infrastructure.
AI is moving beyond the data center
AI is accelerating this transition.
Much of today's AI discussion is dominated by hyperscale data centers, large language models and the enormous infrastructure investments required to support them. Across the industry, vendors are benefiting from unprecedented demand for optical transport, switching, routing, compute, power systems and cooling technologies as cloud providers race to build the infrastructure that will support the first wave of the AI economy.
But Ericsson is looking beyond it.
"While we're talking a lot about the data center today," Tamsons said, "AI will meet the physical world."
That statement reflects a view that is becoming increasingly common among industrial technology leaders. The current data-center boom may be only the first phase of a much larger transformation.
"The first wave now is the data center," Tamsons said. "I'm pretty sure that we will see the next wave around more distributed to the edge."
Based on onsite visits to some of the most advanced industrial and manufacturing deployments in the world, I agree.
The next phase of AI will occur not primarily inside data centers but across factories, warehouses, mines, ports, airports, transportation systems, utilities and critical infrastructure. Intelligence will increasingly be embedded in physical systems, helping coordinate assets, automate decisions and optimize operations in real time. As AI moves into operational environments, networking becomes more important, not less.
The unified infrastructure era
What Tamsons is really describing is larger than a shift in enterprise networking.
It is the gradual collapse of the boundaries that have historically separated telecom, enterprise technology, cloud computing, artificial intelligence and industrial operations.
As AI moves into the physical world, communications networks, cloud platforms, applications, sensors, robotics, automation systems and operational processes increasingly become part of a single interconnected system.
The traditional boundaries that once separated telecom, enterprise IT, operational technology, cloud computing and industrial infrastructure are beginning to dissolve. Technologies that were once deployed, managed and purchased independently are becoming increasingly interconnected as AI, automation and real-time operations create dependencies across the entire stack.
This is the emergence of what I have previously described as the unified infrastructure era: a world in which energy, connectivity, compute, cloud, artificial intelligence, applications and operations increasingly function as components of a single operational system.
The significance of that shift extends beyond technology architecture. It also marks the gradual collapse of the historical boundaries between industries themselves. Telecom is becoming part of enterprise technology. Enterprise technology is becoming part of industrial operations. Cloud infrastructure is becoming part of operational infrastructure. The distinctions that once separated carriers, enterprises, software platforms and industrial systems are becoming increasingly difficult to define.
Viewed through that lens, Tamsons' observation that the network is becoming "part of how operations run" is not simply a statement about enterprise wireless.
It is a description of a much broader transformation. The network is no longer connecting the operation. It is becoming part of the operation.
Thinking beyond the current cycle
One of the more striking aspects of Tamsons' argument is that it runs somewhat counter to the prevailing mood of today's technology market.
Many infrastructure companies are understandably focused on capturing the immediate opportunities created by the current AI boom. Ericsson's perspective is different.
"You need to have a long-term view," she said.
That philosophy has been a recurring feature of Ericsson's culture. The company has often been willing to make strategic bets that appear early, sometimes uncomfortably early, compared with the rest of the market. Its enterprise wireless strategy, its commitment to network APIs and its broader vision of communications as a programmable platform were all built around the assumption that networks would eventually become more deeply integrated into digital operations.
Rather than over-rotating on the immediate requirements or opportunities of today's AI boom, Ericsson is positioning itself for the infrastructure demands that may emerge over the next decade as AI moves beyond data centers and into operational environments.
That is where Ericsson believes the industry's next major opportunity will emerge.
In that future, enterprises may be less concerned with raw network performance and more concerned with whether the network enables operations to continue running reliably, securely and without interruption.
If Tamsons is right, the network will no longer be viewed as a utility.
It will be viewed as operational infrastructure.
And that changes everything.