Written while I was CTO at Orange Business. "Our networks" refers to Orange.
Enterprise connectivity is entering a new phase.
For years, much of the B2B conversation focused on bandwidth, coverage, resilience, and cost. Those still matter. But in the AI era, they are no longer enough to define the opportunity.
What is changing is not only the volume of traffic.
What is changing is the nature of traffic, the shape of enterprise demand, and the strategic role of connectivity in the digital operating model of the enterprise.
This is why I believe the future of B2B connectivity will not be defined by transport alone. It will be defined by connectivity plus intelligence, plus control, plus trust.
The real shift: not just more traffic, but different traffic
The biggest mistake is to think this is only a volume story.
Historically, much of enterprise traffic was relatively predictable and largely human-driven. Users accessed centralized applications, branch-to-data-center patterns were well understood, and connectivity was designed around that model.
That model is changing rapidly.
AI is driving a world with:
- more east-west traffic
- more cloud-to-cloud and data-center-to-data-center exchange
- more interaction between enterprise sites, edge locations, and distributed compute environments
- and increasingly, more machine-to-machine, agent-to-agent, and AI-to-AI communication
We also believe that by 2030, more than two-thirds of traffic on our networks could be AI-driven.
At the same time, we are seeing strong and sustained growth in data-center-to-data-center traffic, as AI workloads, training, inference, and data movement increasingly span multiple cloud and compute locations.
In other words, the network is no longer just carrying user sessions. It is increasingly carrying distributed decision loops.
That is a fundamental shift.
AI changes the requirement for connectivity
When traffic changes, connectivity requirements change with it.
Enterprises now need connectivity that is not only high bandwidth, but also:
- lower latency
- more deterministic
- more resilient
- better able to handle bursty and variable traffic patterns
- more tightly integrated with cloud, security, and observability
This is why AI-ready connectivity is different from traditional enterprise connectivity.
AI is not a side-load on the network. It is becoming one of the forces that shapes network design itself.
Connectivity becomes part of the enterprise operating fabric
As AI scales, the network stops being a passive transport layer. It becomes part of the operating fabric.
Some AI workloads can remain centralized, but many use cases require inference, policy enforcement, and data processing closer to users, devices, factories, branches, and operational sites. That makes edge presence, regional architectures, cloud adjacency, and intelligent interconnection far more important.
This is where connectivity must become more workload-aware.
The future is not simply about moving packets faster. It is about understanding what kind of workload is being supported, what level of performance is required, what policy constraints apply, and what business outcome is at stake.
That is a very different value proposition from selling raw bandwidth.
Where telcos can truly differentiate
With hyperscalers and software-led players moving deeper into the connectivity space, the obvious question is: where can telcos really create differentiated value?
My view is simple: telcos will only capture the AI-era opportunity if they stop competing on commodity transport alone.
If the value proposition is only bandwidth, pressure will continue to increase.
But enterprise customers are not only asking for bandwidth. They are asking for:
- trusted infrastructure
- secure and resilient services
- compliance and accountability
- end-to-end observability
- operational guarantees
- simpler integration across cloud, network, edge, and security domains
That is where telcos can differentiate.
1. Sovereignty
In the AI era, sovereignty matters more.
Enterprises increasingly care about where data is processed, which legal regime applies, who has operational control, and who remains accountable. In regulated sectors and critical industries, this is becoming central, not optional.
2. Hybrid complexity
Most enterprise environments are not cloud-only. They are hybrid, multi-cloud, multi-site, and operationally fragmented.
That complexity is often where customer pain lives.
Telcos are in a strong position if they can simplify that reality across enterprise sites, cloud access, edge workloads, security services, and end-to-end service assurance.
3. Service guarantees and trust
AI-ready trusted connectivity needs more than best-effort transport.
It needs stronger predictability, stronger SLAs, richer observability, and clearer accountability.
In the AI era, speed without trust becomes risk.
The platform shift
One of the most important changes ahead is that connectivity increasingly has to be delivered as part of a platform, not as a standalone product.
Enterprises do not want to stitch together fragmented services across multiple domains of complexity on their own.
They increasingly need a model that brings together:
- connectivity
- cloud access
- security
- observability
- programmability
- control
That is why I believe the future is not just “AI-ready networks.” It is AI-ready connectivity platforms.
As enterprises become more used to platform-based models, real-time interaction, and self-service control across their digital environment, they will increasingly expect connectivity to evolve in the same direction.
What about 6G?
I do believe 6G may magnify the opportunity.
But I do not believe it creates the opportunity by itself.
The real opportunity is already here, driven by:
- AI adoption
- edge architectures
- sovereignty requirements
- software-defined enterprise environments
- the convergence of cloud, connectivity, and digital operations
6G can amplify this by enabling even more dynamic and programmable environments, but only if we already know how to package value beyond connectivity alone.
The future will not wait for 6G to begin.
Final thought
In the AI era, B2B connectivity becomes more strategic, not less.
But the winners will not be the providers who only carry more traffic. They will be the ones who understand that enterprise customers now need trusted, intelligent, AI-ready connectivity across physical and digital environments at the same time.
The future is not connectivity alone. It is connectivity plus intelligence, plus control, plus trust.
And if that future is bigger than any one product or provider, then it also requires a broader industry response.
For all of us in telecom, this is the time not only to compete, but also to collaborate, across connectivity, cloud, AI, security, and platform ecosystems, to build the foundations of the next generation together.
That is the opportunity in front of the industry. And that is what we should be building.
Also published on LinkedIn, where the discussion is.
