- Kelly Ahuja argues sovereignty is about control, not geography
- AI is turning governance into the defining challenge of digital infrastructure
- The future growth opportunity for carriers may be Operational Sovereignty
"Sovereignty is not about a geographic place. Sovereignty is about control and where that control resides."
I've known Kelly Ahuja since 1998.
Back then, Ahuja was one of the key executives behind Cisco's CRS-1, the first terabit router and one of the most significant products in telecom history. At the time, I was running Light Reading. We managed to get ourselves into a fairly spectacular disagreement with Cisco after publishing performance testing that the company did not particularly enjoy reading. (This was back when industry publications were willing to spend $130,000 of their own money to independently test leading vendors' flagship products).
Things escalated.
Engineers argued. Executives argued. Journalists argued. The Wall Street Journal stuck its nose in.
Throughout it all, Ahuja rode above the noise with a kind of Zen-like calm that eventually cooled tempers and restored sanity to the conversation.
Nearly three decades later, he finds himself at the center of another important transition.
As President and CEO of Versa Networks, Ahuja is helping drive a shift that may prove just as significant as the move to IP networking itself. The industry increasingly talks about AI, cloud, security and digital transformation as separate subjects. Ahuja sees something different emerging: a world in which control, governance and sovereignty become the organizing principles of digital infrastructure.
During our conversation this week, he offered perhaps the clearest definition of digital sovereignty I have heard.
That matters because sovereignty is rapidly becoming one of the defining issues of the AI economy. As AI systems become embedded in every layer of digital infrastructure, the critical question is no longer simply where infrastructure resides, but who controls it. The issue extends beyond data sovereignty and cloud sovereignty into questions of policy sovereignty, operational sovereignty and, ultimately, control of the control plane itself.
Those questions formed the backdrop for a conversation that began as a discussion about sovereign secure access service edge (SASE) and ended somewhere much larger.
I had sat down with Ahuja, intending to discuss sovereign SASE, agentic AI, network control planes and the concept I have been calling operational sovereignty, or OpSov. Instead, he spent much of the discussion dismantling the assumptions embedded within those questions.
Like many people in the technology industry, I have tended to think about digital sovereignty as a question of geography. Which country hosts the data? Which jurisdiction governs the cloud? Which government has legal authority over the infrastructure?
Ahuja's argument was different.
Sovereignty, he suggested, is not fundamentally about location.
It is about control.
That distinction matters because AI is changing the architecture of digital infrastructure in ways that make geography increasingly insufficient as a framework.
For the past several years, Versa has been viewed primarily as a SASE company. Yet when I looked at the company's recent announcements, I found myself seeing something else. Sovereign SASE. Sovereign infrastructure. AI-powered operations. Expanded governance capabilities. Taken individually, they looked like product announcements. Taken together, they appeared to describe something much larger.
I asked Ahuja whether Versa was becoming a sovereignty company.
His answer was immediate.
"It's not just sovereign SASE, it's sovereign infrastructure."
That observation became a recurring theme throughout our conversation, and it is a concept that is now echoing throughout the industry.
For decades, networking and security evolved as separate disciplines. Enterprises purchased infrastructure, security products, cloud services and management tools as largely independent components. Classic silos. AI is rapidly breaking those assumptions.
According to Ahuja, enterprises increasingly want infrastructure that allows them to determine not simply where data resides, but who controls it, who governs access to it, where operational authority sits and how policy is enforced.
As the conversation progressed, something else became apparent.
I kept steering the discussion toward agents.
Ahuja kept steering it back toward policy.
Every time I raised autonomous systems, agentic AI or control planes, he responded by talking about governance, business policy, identity, enforcement and observability. We were discussing the same problem from different directions. I was focused on what intelligent systems might do. He was focused on how intelligent systems would be governed.
That distinction is subtle but important.
Much of the current AI discussion focuses on autonomous reasoning, orchestration frameworks, large language models and machine decision-making. Ahuja's concern is different. His concern is how organisations maintain control as those systems proliferate.
That became particularly clear when we discussed what I have previously described as agentic AI chaos theory: the possibility that large numbers of AI agents operating simultaneously may compete for resources, pursue conflicting objectives, create feedback loops and generate unpredictable behaviour.
Rather than dismissing the concern, Ahuja reframed it.
The challenge, he argued, is that AI is introducing an entirely new population of actors into enterprise environments.
According to Gartner estimates he cited, a single human user may eventually spawn dozens or even hundreds of non-human identities.
Those identities will not live in a single place. They will exist across clouds, applications, devices, infrastructure and networks. They will communicate with other agents, access data, invoke services and increasingly make decisions.
The traditional enterprise security model was designed around human users. The emerging enterprise must govern autonomous entities.
That is a fundamentally different problem.
It also explains why Ahuja repeatedly returned to concepts such as zero trust, segmentation, policy enforcement, and observability. His argument is not that enterprises need more AI. His argument is that they need mechanisms capable of governing AI.
The discussion also intersected closely with ideas I have been developing around operational sovereignty. I have argued that the most important sovereignty question is not where infrastructure resides, but whether organizations retain meaningful control over infrastructure, applications, data, AI systems and decision-making.
Ahuja appeared to arrive at a similar destination from a different direction.
He described a world in which infrastructure can be consumed through multiple models. Organizations can consume shared infrastructure, operate dedicated infrastructure provided by a partner, or build and operate infrastructure themselves. What matters is not the physical location of those assets. What matters is who controls them and how governance is enforced.
It is the same question I explored recently in a column asking who controls the control plane of the world.
The implications extend far beyond enterprise networking. They reach into cloud computing, artificial intelligence, telecommunications and critical infrastructure because all of those domains are converging into a single digital infrastructure stack.
Perhaps most significantly, they affect the future role of service providers.
For decades, telecom operators monetized connectivity.
Ahuja believes the next opportunity lies elsewhere.
"The next growth opportunity for service providers is about helping their customers govern, secure and control their digital infrastructure."
That may be the most consequential statement of the entire interview.
What Ahuja describes as the future growth opportunity for carriers is remarkably close to what I have been calling operational sovereignty: the ability of an organization to retain meaningful control over its infrastructure, applications, data, AI systems and decision-making processes. We are approaching the same conclusion from different directions. Ahuja sees it through the lens of carrier strategy and digital infrastructure. I see it as the emergence of a new layer in the technology stack. In both cases, the destination is the same.
Because the future of networking is no longer primarily about transporting traffic. It is about governing increasingly autonomous systems.
And if sovereignty is ultimately about control rather than geography, then the most valuable infrastructure in the AI economy may not be compute, connectivity or even intelligence.
It may be the mechanisms that determine who remains in control once intelligence is everywhere.
Stephen M. Saunders MBE is a communications analyst and USPTO-registered inventor examining how digital infrastructure — 5G, cloud, and AI — is reshaping industry, power and society, as well as underpinning the emerging, ubiquitous global digital economy. As anchor of FNTV and a longtime industry insider, he focuses less on growth narratives and more on execution, risk and how hyperscale technology is distorting markets, governance and society at scale.
Opinion pieces from industry experts, analysts or our editorial staff do not represent the opinions of Fierce Network.