At the FTTH Conference 2026, I had the privilege to moderate a panel on the business-to-business fibre market. One thing was clear: fibre is no longer a background utility. It is becoming a strategic asset — and a prerequisite for the AI economy.
Three forces are driving this shift:
- AI and data-intensive workloads are pushing demand for reliable, high-capacity networks.
- Governments and enterprises are prioritising sovereignty, resilience and regional control.
- And expectations are changing — businesses no longer accept “peak performance” claims; they need consistency they can rely on.
At Opensignal, this shift is familiar. Real-world data shows that performance is not defined by top speeds, but by consistency. That is exactly where fibre is moving into the spotlight.
The market is moving beyond “best effort”
AI is not just increasing traffic — it is raising the bar for network performance.
As Avijit Singh of Colt Technology Services argued, the industry is moving toward a world where some enterprise applications can no longer rely on best-effort connectivity. When AI becomes embedded into real-time decision-making, closed-loop automation and business-critical processes, variability becomes a problem.
For these use cases, deterministic connectivity is no longer optional. Enterprises need predictable bandwidth, stable latency and performance they can trust.
This does not mean every workload will move to private or dedicated infrastructure. Public internet access will still have a role for less critical applications. But the architecture is changing. Enterprises are increasingly likely to combine both models, keeping best-effort connectivity where it is sufficient while using dedicated fibre and private infrastructure where performance assurance matters most.
That is a major commercial and technical shift. Connectivity is no longer simply about access. It is about matching network design to business-critical outcomes.
AI is reshaping traffic flows across the network
The panel also highlighted how AI is changing where traffic goes and where networks need to scale.
East-west traffic between data centres is growing rapidly, driven by replication, resilience and AI model movement. At the same time, inference is pushing workloads closer to users, into metro and edge environments.
The implication is simple: demand is rising everywhere.
And this exposes a blind spot. Investment is pouring into compute and data centres, but the connectivity layer between them is still too often assumed to “just work.” It won’t. If fibre capacity or resilience falls short, the rest of the AI stack underperforms.
The last mile still decides the outcome
That broader investment imbalance came through strongly in the discussion.
There is intense focus today on data centres, AI compute and cloud infrastructure. Yet even the most advanced data centre architecture is not enough if the end-user connection is weak. Denis Teissier of Covage put this: high-performance connectivity only creates value when the end user can actually experience it. Poor last-mile performance, instability, packet loss or lack of resilience will undermine the outcome, no matter how strong the backbone may be.
This is a point the industry sometimes misses. It is easy to focus on the most visible or prestigious parts of digital infrastructure. But the user experience is shaped end to end. Fibre does not create value only between hyperscale facilities. It creates value when it connects data centres, enterprises and users in a way that makes services work without interruption.
That last point is important because it also changes how operators should communicate.
Customers do not buy latency — they buy continuity
Another strong thread in the panel was the gap between how the telecom industry talks and what customers actually care about.
Operators often default to technical language: bandwidth, latency, and resilience. These matter, of course, but they are not how customers think.
As Borislav Tadić of 1&1 Versatel noted, customers are not really asking for a set of technical KPIs. They are asking a simpler question: will my service work when I need it to?
That could mean a cloud application performing reliably, a Point of Sale (POS) terminal staying up, or a video conferencing service holding up without interruption. The value of fibre is real, but the industry still too often falls into the trap of describing it in infrastructure terms rather than business terms and outcomes.
This is one of the most important commercial lessons for the B2B fibre market. If operators want to move up the value chain, they need to translate network excellence into customer outcomes. The conversation should not start with technical specifications. It should start with continuity, reliability, simplicity and most importably building trust.
The SME opportunity is real — but adoption is the real barrier
The panel also brought welcomed realism to the conversation about enterprise digitisation.
In markets such as France, fibre availability and adoption have already advanced significantly, reaching x %. But that does not automatically translate into broad uptake of digital tools, cloud services or AI among smaller businesses. As Denis Tessier pointed out, the challenge for many SMEs is no longer access. It is adoption.
That distinction matters. Many smaller businesses still do not see why they should upgrade or how better connectivity links to practical business value. The barriers are often not technical. They are organisational and human: lack of digital skills, lack of confidence, limited trust and insufficient local support.
This is where the industry needs to be careful. It is easy to assume that fibre coverage solves the problem. It does not. Access is necessary, but it is not sufficient. Without trusted guidance, relevant use cases and clearer articulation of value, many SMEs will remain stuck at the edge of digital transformation, even where infrastructure is already in place.
If Europe wants AI adoption to extend beyond a handful of large enterprises, this is one of the gaps that must be closed.
Security and sovereignty are becoming core connectivity issues
Another major theme was the growing importance of sovereignty and secure-by-design infrastructure.
Phaedra Kortekaas of Eurofiber stressed that the conversation is not about isolation. It is about building a more balanced and resilient digital stack in Europe. That includes stronger control over data flows, more regional infrastructure and more secure connectivity between sites, clouds and data centres.
Quantum security came into that discussion as well. While quantum computing’s full impact may still be ahead of us, the message from the panel was clear: the question is when, not if. For fibre operators, this means security can no longer be treated as an overlay. It needs to be embedded into infrastructure strategy itself.
This raises the strategic importance of partnerships, open access models and cross-border collaboration. If fibre is part of critical digital infrastructure, resilience cannot depend on isolated efforts alone. There is a growing need for the industry to work together more deliberately, particularly where sovereignty, redundancy and national resilience are concerned.
AI may improve telecom — but it may also disrupt it
Perhaps the boldest part of the discussion came around how AI could change telecom operating models themselves.
Borislav Tadić raised a provocative distinction between AI-augmented telcos and AI-native telcos. That is worth taking seriously. Many operators today are using AI to improve existing processes, automate steps and increase efficiency. But the deeper challenge is whether AI could enable new types of operators to build services, operations and customer experience from scratch in a fundamentally different way.
That is not just theory. The panel touched on examples of services being assembled in hours rather than months using agent-driven models. Whether telecom itself can be rebuilt that way remains an open question, but the broader point stands: the risk of disruption is real.
For incumbent and established operators, that means the challenge is not only about deploying AI tools. It is also about rethinking legacy systems, outdated processes and organisational assumptions that slow down change.
Data remains the biggest internal obstacle
When the discussion turned to constraints, one issue stood out above all others: data.
Technology is advancing quickly, and skills can be acquired, partnered for or developed. But fragmented legacy data, weak governance and inconsistent ownership remain major barriers. Without good underlying data, AI cannot deliver at scale. In that sense, the industry’s biggest constraint may not be access to the newest tool, but the condition of the systems and information it already has.
That is an uncomfortable truth, but an important one. AI may help solve some data problems, but only where the underlying foundation exists. If operators want to become more autonomous, more efficient and more outcome-driven, data must become a strategic priority rather than a back-office clean-up exercise.
Fibre’s role is changing — and the industry needs to respond
The next step is not more discussion — it is better measurement and clearer alignment.
If fibre is becoming the backbone of the AI economy, then understanding real-world performance across the entire connectivity chain becomes critical. That means moving beyond theoretical metrics and focusing on how networks actually perform for businesses and users in practice.
At Opensignal, this is exactly where we see the biggest gap — and the biggest opportunity. Operators, infrastructure providers and policymakers that ground their decisions in real-world experience, not assumptions, will be best positioned to deliver on the promise of AI.
The editorial views expressed in this article are solely those of the author and do not necessarily reflect the views of Opensignal.
Opensignal Limited retains ownership of this insight including all intellectual property rights, data, content, graphs & analysis. Reports and insights produced by Opensignal Limited may not be quoted, reproduced, distributed, published for any commercial purpose (including use in advertisements or other promotional content) without prior written consent. Journalists are encouraged to quote information included in Opensignal reports and insights provided they include clear source attribution. For more information, contact press@opensignal.com.
