Time to scale: A telco roadmap for cloud-native success

It’s a wrap on another great DTW event. Heading into the show, Rakuten Cloud SVP and Global Lead Vivek Chadha spoke with Philippe Ensarguet, VP of Software Engineering at Orange and Keith Dyer, editor of The Mobile Network, about the pressure telcos are under to more quickly operationalize cloud-native infrastructure at scale. Let’s dive further into that discussion, the progress operators have made and what needs to happen to close the gap between ambition and reality.

Operators want industrial-grade infrastructure that is scalable, efficient and ready for automation and AI. Progress on this front has been limited, though.

Even the most advanced telcos are operating just 5-10% true cloud-native infrastructure today. With goals to achieve upwards of 60-70% cloud-native by end of the decade, the focus now is on concrete steps that can be taken to meet this ambitious goal.

If progress were just a matter of deploying new tools, operators would already be well on their way. But it’s not as simple as that.

This isn’t a matter of just running more workloads in containers. True cloud-native transformation requires a horizontal model, open ecosystems and new operational practices that span from deployment and lifecycle management to observability and security.

This is the only way they will be able to operate similarly to and compete with their software-native brethren while meeting the scale, sovereignty and service obligations of telecom.

What does it really mean to be cloud-native?

It’s not uncommon to hear someone equate “cloud-native” with containerization alone, but that’s just one technical layer that needs to be considered.

Becoming truly cloud-native means also incorporating:

• Microservices-based architectures
• Day 0/1/2 lifecycle automation (not just deployment but management and updates)
• Service orchestration via Kubernetes-native tooling
• GitOps for deployment and operations (infrastructure + application layers)
• Observability and closed-loop feedback baked into the system

Being cloud-native means having a full operating model that allows telcos to manage services with the same flexibility and tooling that a modern software company would use. Really the only way to achieve benefits like agility, efficiency and automation are rethinking completely how services are designed, deployed and operated.

Let’s explore how technology and telecom organizational workflows are being reconsidered in support of this goal.

Rethinking the operating model: Horizontal platforms and open source

Vertical models are no longer enough.

In the past, telcos would deploy network functions in vendor-specific, siloed stacks on a country or domain basis. This led to a lot of duplication, complexity and lifecycle management fragmentation.

This is an increasingly untenable model for global-tier ones that are prioritizing harmonization, standardization and shared infrastructure across core, edge and RAN deployments.

For instance, Orange’s “Orange Telecom Cloud” (OTC) is built as a common platform (CaaS) for deploying CNFs outside of full vendor-integrated stacks. Importantly, it can be reused across use cases and domains, effectively helping to industrialize CNF deployment and management.

To avoid reinventing the wheel and promote cross-operator alignment, Orange created the Silva open-source project, which is hosted by the Linux Foundation Europe.

Its goal is to define a shared telco infrastructure framework, provide a reference implementation and support validation and lifecycle management of CNFs across vendors and operators. Real telco operator requirements are considered and refined via peer collaboration.

Open source is a major x-factor here as it provides:

• Standardization by creating common frameworks for interoperability.
• Collaboration and innovation which share effort across operators, vendors and integrators.
• Cost-effectiveness by reducing licensing costs and empowering internal teams.
• Accelerated deployment by streamlining development, testing and rollouts.
• Geopolitical flexibility by reducing dependency on restricted tech stacks as the global considerations shift rapidly.

Of course, open source does not equal free. While cost-efficient, sustainable success requires commercial models that support long-term participation and ROI for vendors and operators.

In the case of Rakuten Cloud, we made a deliberate choice to preserve Kubernetes layers and build value on top to futureproof and assure upstream compatibility.

What comes next for telco cloud: Lifecycle, observability, orchestration and security

Even as operators begin to coalesce around clear strategies and open frameworks, they face a number of implementation barriers that must be overcome in the pursuit of scale:

1. Lifecycle orchestration is too complex. Cloud-native transformation doesn’t stop at packaging workloads into containers. Operators must also orchestrate both the infrastructure and the network functions together in a seamless, automated way. That includes managing updates, handling rollback scenarios and ensuring consistency from day zero through day two. Philippe shared how Orange is addressing this via GitOps practices adoption, including tools like FluxCD, to bring version-controlled, closed-loop automation to both infrastructure and service deployments. This shift reflects a broader need to treat telco systems more like software, with fully automated lifecycle operations that can scale across diverse environments.
2. Observability is critical but has never been more challenging. Legacy networks were easier to monitor because each function was monolithic. In cloud-native environments, those functions are broken into dozens or more microservices that are spun up and down rapidly. Without deep, Kubernetes-native observability, operators risk losing sight of how services are performing, where bottlenecks are forming and whether SLAs are being met. To support automation and service assurance, observability has to capture real-time telemetry across layers and expose meaningful insights.
3. Security poses a major operational challenge. As operators adopt open infrastructure models, they need better transparency into what’s being run. As Philippe highlighted, software bills of materials (SBOMs) have emerged as a baseline requirement for overcoming the near-impossible task of assessing vulnerabilities and managing risk across supply chains. After all, operators can only protect what they can see and understand. We have moved to an era where security must be at the forefront of design, and built into how software is developed, validated and deployed.
4. Multi-network complexity compounds difficulty. Managing a mix of 4G, 5G, Wi-Fi, fixed access, and edge increases the burden on orchestration frameworks, which now must accommodate different topologies, performance needs and technology stacks. The path forward is to unify these environments within a shared platform with careful abstraction and highly configurable orchestration logic, and support from new operational models and better cross-team coordination.

These challenges reinforce the earlier point that telco transformation can’t just be about adopting new tech. The road to success requires building operational support to manage dynamic, distributed systems at scale. All the while, telcos are keeping an eye on an important follow-on goal: establishing AI-driven autonomous operations.

AI only as effective as the cloud-native foundation supporting it

Perhaps no topic received greater airtime at DTW this week than AI.

As operators work toward AI-driven operations, they recognize it can’t just be layered on top of legacy environments.

The same cloud-native infrastructure we’ve discussed here to enable scalable, automated network function deployment will also ultimately support AI workload execution and management.

Telcos working toward fully cloud-native for CNFs need to bear in mind that the same framework will also support AI inference workloads and model orchestration, revealing a dual-importance for cloud-native journeys.

As Vivek pointed out, AI is already being applied to new types of observability with energy consumption joining logs, metrics, events and traces as a fifth pillar.

For its part, Rakuten Cloud has begun using machine learning to optimize energy usage at the infrastructure layer to reduce power draw without degrading service quality. This is particularly impactful in telecom, where always-on networks have previously left little room to conserve energy through conventional methods.

A boost from AI may see operators finally starting to identify and adjust system behaviors in ways that were previously impossible.

Beyond infrastructure, AI is being applied to product development and service lifecycle planning to help business users pose operational questions and receive end-to-end impact assessments from generative tools.

By connecting AI to the same platform layer that supports network workloads, telcos can create a shared operational fabric that supports both human and machine intelligence.

Transformation is underway, but the clock is ticking

Cloud-native infrastructure is a foundational requirement for establishing telco capabilities that compete with software-native businesses. It is the only way to achieve scalable services and AI integration, while managing complexity without performance compromise.

As we have discussed here today, the shift will not only be technical but also operational and cultural, requiring shared platforms, new practices and close cross-industry collaboration.

Success is within reach but the window to make this transition is narrowing.

Operators have only a few short years to move from early deployments to full-scale execution before the demands of 5G-Advanced, 6G and AI-first operations exceed what legacy models can support.

What steps is your organization taking to close the gap between where you are today and where you want to go? How are you plotting an urgent path to cloud-native?

Let us know in the comments and be sure to check out our Zero-Touch Live episode featuring Keith Dyer of The Mobile Network, Philippe Ensarguet of Orange and Vivek Chadha of Rakuten Cloud.