Building and Scaling AI the Cloud Native Way at Singtel

At FutureNet Asia 2025, Vinod Joseph, Vice President for Cloud, AI and Enterprise Architecture at Singtel Group, shared a deep dive into how telcos can scale AI in a cloud native manner. His talk moved beyond the buzz around generative AI and looked at the infrastructure and operational realities required for long-term success.

Vinod positioned the industry as moving into a second phase of AI adoption. The early phase focused on agents, copilots and low-code platforms. The next phase, however, demands robust systems for managing data pipelines, training and fine-tuning models, monitoring model performance and deploying AI in production environments at scale. He stressed that this shift requires not only flexible cloud environments, but also consistent engineering practices and strong governance frameworks.

A core theme of the session was the importance of avoiding proprietary lock-in. Vinod argued that as AI workloads grow, organisations need the freedom to deploy where it makes sense, whether on-premises or on public cloud, while maintaining agility and operational consistency. Kubernetes featured strongly as a foundational platform for AI workloads, offering orchestration capabilities and portability across environments.

Three open-source frameworks were highlighted as central to Singtel’s approach. Kubeflow supports the orchestration of AI and machine learning pipelines, handling key stages from model training to promotion into production. Ray helps distribute compute workloads across GPUs and servers, enabling efficient training of large-scale models where data and model components cannot fit on a single device. MLflow, meanwhile, provides experiment tracking, model registry and deployment management, simplifying lifecycle operations and improving observability.

Vinod stressed that scaling AI requires more than computational power. Efficient data handling, reproducibility, experiment lineage and reliable recovery from failures are just as important. As organisations accelerate AI adoption, these capabilities become essential not only for performance, but for cost control. Open-source tooling, he argued, is becoming increasingly competitive and offers a viable way to balance capability with economic scale.

Singtel’s perspective reflects a growing maturity in telecom AI strategy. The focus is shifting from exploration to industrialisation, from early pilots to repeatable and governable systems. With cloud native architectures, distributed computing and open frameworks at the core, the goal is to build platforms that can scale flexibly, avoid dependency on any single vendor and support the next generation of AI-driven services.

The full presentation is available below for anyone who would like to watch it:

Related Posts: 

• The 3G4G Blog: What is Cloud Native and How is it Transforming the Networks?
• Operator Watch Blog: 6 Mobile Operators in the Global 100 Ranking of the World’s Most Sustainable Companies
• Operator Watch Blog: Singtel Surpasses 95% Nationwide 5G Standalone Coverage in Singapore
• Operator Watch Blog: Singapore is a 5G leader in Southeast Asia
• Operator Watch Blog: Singtel shows the power of Standalone 5G
• Private Networks Technology Blog: Singtel’s 5G Powers the Transformation of Singapore’s Tuas Mega Port
• Private Networks Technology Blog: 5G Private Networks Driving Industry Transformation in APAC
• Operator Watch Blog: Cloud Native Progress and Pain Points According to Orange
• The 3G4G Blog: Cloud Native Telco Transformation Insights from T-Systems
• Operator Watch Blog: How AI Is Reshaping Network Operations at Deutsche Telekom
• Operator Watch Blog: Singtel Surpasses 95% Nationwide 5G Standalone Coverage in Singapore 

How AI Is Reshaping Network Operations at Deutsche Telekom

Michal Sewera, an experienced technology leader at Deutsche Telekom Group (generally written as TDG which stands for 'Telekom Deutschland GmbH'), recently offered a rare behind-the-scenes view of how AI is being used to manage and optimise telco cloud operations. As the head of TDG’s cloud-native 5G core DevOps team, he has led the shift to a new operating model built on cloud-native principles, automation and AI.

Presenting at the FutureNet World conference in London on 7–8 May 2025, Michal shared how TDG’s journey to cloud-native began with the realisation that cloud is not simply about virtualisation or containers. The real transformation lies in a fundamental change in architecture and operations. Moving to a GitOps operational model with declarative deployments and a concept of desired network state has allowed TDG to move from infrequent bulk updates to continuous, incremental changes. In this new approach, change is no longer an exception but an asset.

However, this shift comes with its own challenges. Cloud-native telco systems are composed of highly distributed microservices, open-source components and loosely coupled layers. This creates what Michal refers to as the butterfly effect, where even a small change can lead to unexpected consequences elsewhere in the system. Traditional approaches to validation, configuration and assurance are simply no longer sufficient.

To address this, TDG has integrated AI tools across all stages of the network lifecycle: development, rollout and operations. In the development phase, TDG uses an AI-based validation framework that collects data from across the application, platform and infrastructure layers. It analyses complex interdependencies using pattern recognition across 3GPP signalling, KPIs, logs, Kubernetes, CNIs and service mesh. This approach replaces traditional regression testing with intelligent analysis that highlights functional issues and pinpoints root causes early in the pipeline.

During rollout, the AI-powered Network Configuration Co-Pilot supports configuration changes across distributed clusters. The tool goes well beyond Git automation bots, using a mix of reusable configuration patterns, chat-based interaction with embedded vendor knowledge and natural language integration with systems like Kubernetes. This allows engineers to handle the massive complexity of telco configurations more efficiently and with greater confidence.

In live operations, TDG employs a combination of active and passive monitoring across its Platform as a Service layer. Probes and telemetry continuously monitor performance while AI-driven root cause analysis tools detect anomalies and correlate them with platform and network data. This enables early detection of degradation and supports predictive fault analysis. TDG also applies AI to canary testing and deployment. New releases are gradually introduced in production environments under close AI-assisted monitoring, allowing issues to be caught before full rollout. This model is a marked departure from the old reliance on staging environments and lab testing.

TDG’s new operational model, grounded in GitOps and driven by AI, offers a compelling example of how operators can adapt to the complexity and speed of change in cloud-native environments. The shift transforms telecom networks from silent, black-box systems into transparent, data-rich platforms where actionable insight can be extracted and acted upon in near real time.

Michal’s insights make clear that AI is not an optional add-on in this new environment. It is a fundamental enabler that allows the telco cloud to scale, evolve and remain resilient. For operators looking to modernise their networks, TDG’s experience offers valuable lessons in how to harness automation and intelligence to meet the demands of the future.

You can watch the full video of his talk below:

Related Posts: 

• The 3G4G Blog: Cloud Native Telco Transformation Insights from T-Systems
• The 3G4G Blog: What is Cloud Native and How is it Transforming the Networks?
• Operator Watch Blog: Cloud Native Progress and Pain Points According to Orange
• The 3G4G Blog - Lessons from ANRW ’24: AI and Cloud in 5G/6G Systems
• The 3G4G Blog: Disaggregation of 5G Core (5GC) Network
• The 3G4G Blog: Reigniting Growth in the Telecom Industry with AI and Cloud
• The 3G4G Blog: Transitioning from Cloud-native to Edge-Native Infrastructure
• Free 6G Training: Cloud Native Architecture & Migration Towards 6G from B6GS 2023
• Free 6G Training: SK Telecom Claims ‘Bare Metal-Based Cloud-Native 5G Core’ as Foundation for 6G Evolution 

How O2 Telefónica is Redefining the Network with NaaS and Open APIs

At O2 Telefónica Germany, the “network of the future” is no longer a distant vision, it’s becoming an operational reality. Under the leadership of Matthias Sauder, Director Networks, the operator has made substantial strides in transforming its end-to-end infrastructure to meet the demands of a highly agile, programmable, and customer-focused future.

At the FutureNet World conference in London (7–8 May 2025), where leading telco and tech stakeholders gathered to explore network automation and AI in telecoms, Sauder shared how O2 Telefónica’s evolution journey spans radio, transport, core, data centres, and cloud landing zones, all underpinned by a strategy that prioritises automation, flexibility, and openness.

The transformation began with a foundational goal: enhancing radio network quality. Once considered the underdog in a three-operator market, O2 Germany set out to radically improve performance by embracing agile practices and reshaping internal structures. Inspired by the Spotify model, the company introduced agility not just in project management but also in technical delivery. This included frequent software release cycles for radio, positioning itself as a global leader in rapid deployment and continuous integration.

A key initiative known as Tech Strategy 25 laid the foundation for modernising radio, transport, and core networks. Today, over 80 percent of the strategy has already been executed. With a largely cloud-native core in place, O2 Telefónica is among the pioneers of this architectural shift. Its collaboration with Ericsson produced one of the world’s first cloud-native digital cores, while a parallel effort with Nokia deployed core services for one million users in a public cloud environment.

The rationale behind both cloud-native and public cloud approaches is clear. Legacy architectures no longer support the operational agility or cost efficiency needed in today’s competitive telecom landscape. Cloud-native systems enable advanced capabilities such as continuous integration, continuous delivery, and seamless in-service software upgrades (ISSU). O2 Telefónica has shown these upgrades can be executed without disrupting live customer services, challenging the long-standing perception that such practices are too risky for telco-grade reliability.

Beyond infrastructure, the company’s future network model hinges on the integration of open APIs and Network-as-a-Service (NaaS) capabilities. These aren’t abstract concepts, they’re practical tools enabling agility, programmability, and new revenue streams. Open APIs expose network functions to external developers and partners, unlocking opportunities for co-creation and monetisation that were previously out of reach.

This openness also extends to industry partnerships. A standout example is the company’s collaboration with Siemens, which now leverages O2's slicing capabilities to deliver tailored network services to its own customers. These kinds of arrangements demonstrate how NaaS, built on secure and standardised APIs, can unlock vertical-specific innovation.

But transformation isn’t just about technology, it’s also about mindset and culture. Simplifying and standardising network configurations (for example, reducing radio setups from over a hundred to just two) and promoting a service-centric approach are part of a broader shift. The focus is firmly on use cases and customer value, avoiding the trap of deploying technology for its own sake. Every new system or tool must demonstrate end-to-end value.

O2 Telefónica also recognises that data, rather than AI alone, is the foundation of intelligent automation. Without a robust data strategy, ambitions around AI, closed-loop automation, or service orchestration are unlikely to succeed. The company’s investment in OSS transformation and data-driven operations is laying the groundwork for intelligent networks that can scale, adapt, and optimise in real time.

As the line between network and IT continues to blur, O2 Telefónica is aligning its BSS, OSS, and IT systems with its network strategy. This integrated approach supports holistic innovation and positions the company to deliver services with faster time to market and greater cost efficiency.

The transformation journey shared by Matthias Sauder is more than a technical roadmap, it’s a call for industry-wide disruption. With revenues flat and operational costs rising, embracing NaaS, open APIs, and cloud-native infrastructure is no longer optional. It’s the only viable path for telcos to stay competitive, innovative, and relevant in a software-defined, platform-centric future.

Sauder’s full presentation at FutureNet World provides deeper insight into this journey. You can watch it below:

Cloud Native Progress and Pain Points According to Orange

Four years ago, the idea of cloud native in telecom was mostly aspirational—an ambitious leap from legacy architectures toward agility, automation, and scale. Today, while the journey is well underway, the destination is still far off.

At the recent Telco to Techco session, Philippe Ensarguet, VP of Software Engineering at Orange, took to the stage to assess the industry's real progress—and expose where it’s still struggling.

Telcos Are Still Caught Between Two Worlds

Many telecom functions are now containerised, but Philippe makes it clear: that doesn’t mean they’re cloud native.

“The ‘C’ in CNF must stand for Cloud Native, not just Container.”

Cloud native isn’t just a new way to package software—it’s a new way of building, deploying, and managing it. And that shift is proving to be far more complex than simply adopting Kubernetes or moving to public cloud.

Legacy virtualised network functions (VNFs) weren’t built for the dynamic, distributed nature of cloud platforms. Trying to retrofit them often results in complexity without the expected benefits.

What’s Working: Areas of Maturity

Despite the challenges, some progress is undeniable:

✅ Infrastructure Automation

Telcos like Orange have built robust cloud native platforms based on open technologies. The ability to scale infrastructure efficiently and reliably is now a reality.

✅ GitOps & Lifecycle Management

Cloud native lifecycle tooling—especially GitOps—is maturing. Orange, for example, manages diverse vendors through a unified GitOps-based integration platform called Network Integration Factory Tooling Zone.

✅ Open Source Participation

Open ecosystems are no longer optional—they’re essential. Orange is actively involved in Project Sylva (under Linux Foundation Europe) to define open, telco-grade cloud infrastructure.

What’s Still Holding Us Back

🛑 Skills Gap

Cloud native demands both hard skills (microservices, APIs, automation) and soft skills (agile mindsets, DevOps culture). These aren’t always easy to find—or to develop—in traditional telco teams.

🛑 Vendor Maturity

While some vendors are rearchitecting their software, others are just lifting old VNFs into containers. Philippe emphasises that cloud native transformation must go deeper.

🛑 Distributed Complexity

Managing services across private cloud, edge, and public cloud creates orchestration challenges. Real-time and asynchronous network functions must coexist—something telcos still struggle with operationally.

Why Cloud Native Still Matters

Despite the friction, the reasons to go cloud native haven’t changed. If anything, they’re more relevant than ever:

• Scalability for on-demand growth
• Agility for faster feature rollout
• Resilience for improved service continuity
• Efficiency to reduce infrastructure and operational costs
• Innovation via open APIs and open source ecosystems
• Multi-cloud flexibility and reduced vendor lock-in

Final Word

Philippe’s closing message was both grounded and optimistic. Yes, the journey is complex and sometimes slow—but cloud native is no longer a buzzword. It’s becoming the backbone of the telco techco transformation.

“The next few years will be about closing the gap—not just between CNFs and legacy systems, but between ambition and execution.”

A detailed article is available on Mobile Europe website here. The video of the conversation is embedded below:

Related Posts

• Operator Watch Blog: Orange Africa & Middle East (OMEA) to Reach Net Zero by 2040
• Operator Watch Blog: Energy and Carbon Footprint Reduction are a Priority for Orange
• The 3G4G Blog: What is Cloud Native and How is it Transforming the Networks?
• The 3G4G Blog: Transitioning from Cloud-native to Edge-Native Infrastructure
• The 3G4G Blog: Reigniting Growth in the Telecom Industry with AI and Cloud