By Martin Reilly, Head of International MOFN, Ciena
The world’s largest cloud providers, also known as hyperscalers, are accelerating a global infrastructure build-out to meet surging demand for resilient, low-latency services. To handle this surge, these companies have invested heavily in submarine cables. A key example of this is Meta’s Project Waterworth, a $10 billion, 50,000km submarine cable system intended to link five continents and employing deep-water routing up to 7,000m.
At the same time, hyperscalers are expanding their data center footprints – moving from urban hubs to rural and strategically advantageous locations. However, the investments being made into subsea cables only deliver data to shorelines. To reach inland data centers, hyperscalers must overcome regulatory and logistical hurdles. And while hyperscalers excel at building out their undersea infrastructure, terrestrial networks present unique challenges, leaving a critical opportunity for telcos to step in and bridge the gap.
This is where telecommunications service providers (telcos) come in, partnering via Managed Optical Fiber Networks (MOFNs) to close this gap securely and at scale.
Connecting Shore to Inland
Investment in new connectivity routes and services is booming. According to TeleGeography’s 2025 State of the Network report, content and cloud providers now account for more than 70% of all international bandwidth usage, a dramatic shift from just a few years ago when internet wholesalers still dominated most routes. Since 2019, global bandwidth demand has surged – nearly tripling – especially along core corridors connecting major North American, European, and Asian data center hubs. However, since most data centers are located well inland, sometimes by thousands of kilometers, connecting them to the heavily invested in subsea cables is no easy feat.
This becomes increasingly challenging as, in many regions, securing dark fiber – the optical fiber cables already installed but not currently in use – is difficult, if not impossible, due to regulatory or market constraints. In North America, for instance, laying a cable from Los Angeles to New York would require navigating a maze of regulatory approvals across states, municipalities, and private lands, each adding layers of bureaucracy, cost, and delays. In regions like Africa and Asia, hyperscalers face even greater obstacles, with stringent regulations often barring them from building, owning, and operating terrestrial networks. Moreover, licensing to own or operate terrestrial networks can be a lengthy, uncertain process.
The Hyperscaler-Terrestrial Gap
While hyperscalers are relatively new to owning their own terrestrial networks, telecommunications service providers (telcos) have decades of experience in this field. Telcos have developed the local expertise and regulatory insight needed to connect submarine cables to inland data centers, uniquely positioning them to support hyperscalers’ infrastructure build out. Telcos, with their extensive terrestrial fiber holdings, including large reserves of dark fiber, can remove many obstacles for hyperscalers – providing different options such as leased-capacity or wavelength services. While hyperscalers may face barriers to acquiring dark fiber directly, for instance, telcos already have extensive dark fiber assets that they can activate or expand to support new connections for their customers.
Moreover, telcos often maintain well-established operations for managing emergencies such as physical fiber cuts, natural disasters, or unplanned outages. This readiness not only reduces risk but accelerates deployment timelines by relying on tested procedures and relationships with local authorities. Additionally, telcos’ existing fiber networks typically span commercial, industrial, and rural regions.
That said, leasing dark fiber or basic capacity can be a slow and complex route, with long negotiations, licensing issues in certain regions, and can limit the ability to customize a network to hyperscalers’ technical needs. All of this can delay market entry and restrict hyperscalers’ ability to scale quickly. This is where Managed Optical Fiber Networks (MOFN) is emerging, enabling telcos to monetize their assets by supporting hyperscaler growth in a highly collaborative way.
A collaborative path forward
MOFN is a service offering whereby a service provider builds a dedicated fiber network for a customer, such as a hyperscaler and provides it to them as a managed service. MOFN has been around for a long time, but the business model is experiencing a renewed momentum and increased traction as hyperscalers expand operations and seek efficient ways to interconnect their sprawling infrastructure. The MOFN model also allows telcos and hyperscalers to work together to secure dedicated, and often customized, high-performance connectivity. MOFN arrangements between telcos and hyperscalers typically see the telco designing, building and operating the network as a managed service. This is especially valuable in regions where regulatory restrictions make it challenging, or even impossible, for hyperscalers to build and operate networks themselves.
Crucially, MOFN is a business model that benefits both the traditional telco and the large scale cloud providers. For telcos, it provides these traditional service providers with incremental and reliable revenue, as well as allows them to form a long term partnership with a hyperscaler over time. For hyperscalers, this model gives them speed and scale at a low cost and takes away the operational burden of running a network, as telcos already have the right infrastructure and ecosystems in place. Telcos will already have an established ticket process, for instance, as well as the management of fiber cuts and potential relocations. Telcos are built to navigate changes to their networks, planned or not, and already have fully formed ecosystems in place to adapt quickly.
Telcos also have pre-established relationships with technology vendors, who act as critical facilitators in the MOFN ecosystem. In this collaborative model, hyperscalers outline their needs for performance and expansion, while telcos showcase their infrastructure capabilities. Vendors then position the right technologies to bridge these objectives, creating win-win scenarios that accelerate deployment and innovation.
Evolving MOFN models
When it comes to MOFN models, typically the telco owns and maintains the infrastructure through strict service-level agreements, and commercial terms often combine one-time fees for fiber access and installation with recurring charges for management, maintenance, and capacity upgrades.
This is not always the case, however, as the MOFN model is flexible, and hyperscalers are exploring the many different ways the model can serve them and their specific needs. For instance, some hyperscalers opt for dedicated, end-to-end fiber networks delivered as a managed service, and purchase the entire system all from day one. Others initially just take advantage of managed wavelength or capacity services, which allow them to scale up quickly as demand grows. Increasingly, there are also hybrid models, in which the service provider manages the underlying optical line system, but the hyperscaler retains control of certain network elements. This model offers both flexibility and future-proofing, which will be key as requirements evolve.
The AI Opportunity
A MOFN business model allows telcos to play a critical role in the deployment of existing and new cloud services, such as artificial intelligence (AI). This is because the rising demand for AI workloads, particularly training and inference for huge models, generates enormous data flows across and within data centers, in turn spurring exponential Data Center Interconnect (DCI) growth. In fact, recent research found that data center experts predict at least a 6X increase in DCI bandwidth demand over the next 5 years. This is where MOFN comes in.
MOFN models leveraging the very latest network technologies and services provide reliable, secure, and scalable connectivity for Data Center Interconnection (DCI) services over vast distances, overland and undersea, in a cost-effective and sustainable manner. As MOFN can be customizable, it also allows cloud service providers and hyperscalers to purpose-build the design of the MOFN model they’re harnessing to connect their data centers and AI factories.
The rise of intelligent networking
In parallel, AI-driven software tools will become indispensable for improving network management. These technologies enable telcos to optimize performance in real-time, anticipate and address potential issues, helping to reroute traffic, ensure redundancy and manage the complexity of expansive networks with greater intelligence.
As these networks expand in scale and complexity, telcos are increasingly relying on AIOps (Artificial Intelligence for IT Operations) to manage them efficiently. These platforms harness AI, machine learning (ML) and advanced analytics to ingest diverse data, detect anomalies in real time, and forecast outages before they happen. For hyperscalers, this means greater reliability, faster response to incidents, and networks that can adapt dynamically as business needs change.
As the demand for cloud services accelerates and new markets come online, collaboration between telcos and hyperscalers will be vital in closing global connectivity gaps, both undersea and inland. MOFN enables hyperscalers to access new markets quickly, without the burden of building terrestrial networks, while offering telcos new revenue streams and partnership potential. This model is a clear win‑win, strengthening competitive positioning across an ever‑evolving landscape.
