• Deutsche Telekom’s AI revenue and cost targets dominated a week of bold operator financial commitments around AI
• Sovereign AI infrastructure, autonomous network trust frameworks and enterprise AI deployment were also prominent themes

Operators and their suppliers spent the week setting out the financial stakes of AI adoption, with Deutsche Telekom putting firm numbers on expected revenue gains and cost savings at a dedicated AI investor day. Elsewhere, themes of sovereign AI infrastructure, trustworthy autonomous systems and the challenge of translating AI pilots into production deployments ran through announcements from across the industry. 

So here are the main AI-native telco developments from the seven days that ended on Friday 9 October…

Tim Höttges, the ambitious and outspoken CEO of Deutsche Telekom (DT), used the telco’s first AI Investor Day (Monday 5 October) to highlight the potential of driving near-term incremental AI-related revenue opportunities, predicting his company will generate an additional €1bn-plus in revenues outside the US by 2030 as a result of such investments.  DT, one of the AI-native telco frontrunners, is already generating revenues from AI services developed for its B2B/enterprise customers outside the US and these are set to increase from an expected €250m this year to €800m in 2030, as previously reported. But Höttges went a step further in his investor day presentation, noting that the overall opportunity for DT’s AI-related service revenues is set to exceed €1bn in 2030 (outside the US), and could even be much greater than that figure. The CEO stated: “€1bn in extra revenue [outside the US by 2030] is something we have in mind. To be honest, this is a conservative number. I do not want to commit to a bigger number because you will not believe it. So we go with €1bn, but our internal plannings are much more optimistic than what we have laid out. Let’s start with €1bn and then we will see. But the opportunities along that value chain are unique for a telco.” He concluded: “These are our audacious targets going forward. And by the way, that doesn’t mean that they are visionary; they are just implemented in our planning process, which we are working out right now.” Read the full story.

Deutsche Telekom’s numbers will be of great interest to other telcos, all of which are trying to figure out what kind of impact their investments in, and use of, AI will have on their financial performance and operating efficiency. Other operators are starting to share such information: Globe Telecom in the Philippines, for example, recently reported that, according to its internal calculations, it has achieved a 2.8x return on its AI investments to date. Both Deutsche Telekom and Globe are among the 56 operators around the world that are being tracked and analysed for TelecomTV’s AI-Native Telco Index reports, the first of which has just been published – see New report identifies the AI-native telco frontrunners.

A1, the national telco in Austria and the flagship operation of central European telco A1 Telekom Austria Group, has expanded its relationship with IBM to offer AI services “on sovereign infrastructure for companies and organisations” in the central European country, the telco has announced (in German). The technology foundation for A1’s Exoscale cloud platform is IBM watsonx, which comprises “watsonx.ai for the development and use of generative AI, watsonx.data for data preparation and management, watsonx.governance for AI governance and compliance, and watsonx Orchestrate for agentic AI and automation scenarios. Thanks to the open architecture approach, companies can flexibly integrate and operate different AI models – from IBM models to open-source models and their own custom models,” noted A1. The telco’s aim is to provide Austrian businesses with a “powerful enterprise AI platform on European cloud infrastructure… a robust and trustworthy foundation for deploying artificial intelligence, where data sovereignty, transparency and regulatory requirements are considered from the outset.” According to A1, the service “supports companies in implementing regulatory requirements, such as those related to the EU AI Act. With watsonx.governance, AI applications can be documented, monitored and controlled throughout their entire lifecycle. This creates transparency, traceability and trust in the productive use of AI. Offerings range from trusted AI assistants and document solutions to AI-powered productivity applications, governance-as-a-service offerings and industry-specific AI services,” added the telco. Martin Resel, currently deputy CEO and chief customer officer (CCO) for the enterprise sector at A1 (though he is leaving shortly – see below), noted: “Our customers expect high-performance AI solutions that meet the highest standards of security, compliance and data sovereignty. By combining Exoscale’s European cloud infrastructure with IBM’s AI expertise, we are creating an attractive offering that enables companies to implement their AI strategy confidently and sustainably.” Sovereign services and strategies will be the focus of TelecomTV’s upcoming Digital Sovereignty Forum, which will be held in London on 2 December.

Mindful that businesses need as much help as possible to get the greatest benefits from the use of AI, Indosat Ooredoo Hutchison (IOH), one of the largest telcos in Indonesia and home to one of the most ambitious AI-native telco strategies in the world, has teamed up with security tech specialist F5 to develop the Indosat Intelligence Suite, which provides enterprises with “a trusted and governed environment that simplifies how businesses access and deploy AI, without requiring them to build and manage complex infrastructure themselves.” The telco notes in this announcement that the platform “builds on Indosat’s full-stack AI ecosystem in Indonesia, spanning energy, datacentres, compute, connectivity and AI models” to bring “multiple AI models and infrastructure capabilities together through a single interface, giving businesses greater visibility and control over how they use AI. It is designed to help organisations select and deploy the intelligence suited to their specific workloads while maintaining the security and governance required for production environments,” added the operator. IOH’s president, director and CEO, Vikram Sinha (whose five-year contract extension has just been approved by the company’s shareholders) stated: “As AI moves from experimentation into everyday business, companies need more than access to models, they need a trusted way to put intelligence into production. With the full stack AI as the foundation, Indosat Intelligence Suite will become an intelligence factory for businesses, helping businesses access the right intelligence for their needs and scale it securely across the region.” IOH noted that the launch “marks the next step in Indosat’s AI journey: Building not only AI capacity for Indonesia and the region but the way to use it. By connecting infrastructure, models and enterprise demand, the Intelligence Suite can help businesses put growing AI capacity to work and support wider AI adoption and economic value across the region.” IOH recently helped to form Zankore, an AI infrastructure and services joint venture with Nvidia, Nokia and Ooredoo Group that recently secured $3.1bn in debt funding.

As with many telcos that are ahead of the pack with their AI-native telco plans, IOH is working very closely with Nvidia, which is very keen to see AI hardware and software integrated into communications network infrastructure around the world, and indeed IOH has featured in Nvidia’s latest blog about telco AI strategies. The AI tech giant is extolling the virtue of open models for telcos that want the “ability to trust, control and customise AI across their most critical workloads – from autonomous networks to customer care”. Nvidia highlights how the adoption of open models is helping Japan’s SoftBank Corp, US telco giant AT&T and IOH to get the most out of AI. As you would expect, the examples all point to how important a partner Nvidia is to these operators (as it is to an increasing number of telcos), but the blog is still instructive as to how and why certain approaches are paying off for the telcos. For example, Andy Markus, AT&T’s chief data and AI officer, notes: “At AT&T, we believe the future of AI is not about choosing a single model; it’s about intelligently matching every workload to the right combination of performance, cost and control. Open models are essential to that approach and, with Nvidia, we can bring that model-choice strategy into production with the scale, reliability and governance our business requires.”

The TM Forum has partnered with Accenture to launch a new industry benchmark aimed at certifying trusted AI. The aim is to “help the telecom industry to deploy autonomous operations with verifiable evidence that agent-to-agent interactions and decisions are trustworthy and can deliver business outcomes at scale,” noted the forum. The move comes at a time when the trustworthiness of frontier AI models is being called into question due to multiple instances of unexpected and insecure behaviour: Recent reports claim AI agents have hacked a number of institutions and companies, despite safeguards put in place by AI giants such as OpenAI and Anthropic. Just this week, OpenAI apologised to the Australian government after a rogue AI model hacked some of its systems. As usage of models grows, and telcos look at how to introduce more AI and automation into their networks, they are under “exceptional pressure to demonstrate business results”, which means “trust is becoming the pre-requisite for delegating decisions and actions to autonomous systems”, the TM Forum noted in its announcement. The industry body has worked with Accenture to develop its Trustworthy AI Governance Certification, which it claims will provide a practical measure of the maturity, workforce readiness and operational accountability needed to confidently scale AI and autonomy. “Trust enables autonomy to scale, but trustworthiness only becomes operational when every principle is supported by continuous, verifiable evidence. This programme gives the industry a common blueprint to turn trust from a promise into a measurable capability. TM Forum brings the industry together to define the standards, while Accenture brings the experience of putting them into practice. Together, we’re creating the foundation for the autonomous enterprise,” stated Guy Lupo, EVP of trustworthy AI and data at TM Forum. As part of the initiative, CSPs will be able to: Measure trust consistently and reduce complexity with a common framework; accelerate AI adoption through best-practice guidance; benchmark trust maturity and expose critical gaps using AI trust level validations; and build enterprise-wide capability and workforce readiness through training and certification. CSPs, vendors, hyperscalers, consultancies and the wider ecosystem will collaborate to co-create code-first solutions through TM Forum’s AI-Native Collaboration Studios, the industry body added.

Singtel’s sovereign cloud unit RE:AI, part of the telco’s Digital InfraCo division, has teamed up with the Singapore Institute of Technology (SIT) to help enterprises translate promising AI pilots into production deployments, something that Singtel notes is a challenge because off-the-shelf AI solutions do not always meet the specific business needs of enterprises. “Through the partnership, enterprises participating in the Digital InfraCo’s Centre of Excellence (CoE) for Applied AI with Nvidia will collaborate with SIT students, researchers from the SIT x Nvidia AI Technology Centre (SNAIC) and the Nvidia AI Technology Centre (NVAITC) to develop AI applications for industry use cases,” noted Singtel in this announcement. “Enterprises and government agencies will be able to co-create purpose-built AI solutions, leveraging RE:AI’s sovereign AI cloud infrastructure and network of technology partners with SNAIC’s applied research expertise, industry collaborations and AI education capabilities,” it added. RE:AI was in the news recently for helping Singtel to become one of the first telcos to launch an AI token-as-a-service (TaaS) offering.

Jeong Jae-heon, the CEO at South Korean telecom operator and self-styled ‘AI company’ SK Telecom (SKT), has presented the new SKT Growth Gear Strategy to the company’s workforce, stating that SKT’s key goal is to position itself as the “gateway to the token economy” by “organically connecting key businesses – telecommunications, AI services and AI infrastructure – with AI technology and AX (AI transformation) as common driving forces”. The CEO noted in this announcement (in Korean) that the “axis of AI competition is expanding beyond models and infrastructure to an ecosystem that stably generates, supplies and utilises tokens… this represents a new growth opportunity for SKT, which has consolidated its telecommunications and AI capabilities.” As part of the new plan, SKT is developing a new business called Token Gateway, which “acts as a control centre that manages AI usage patterns for companies and customers and connects them to optimal AI models and infrastructure.”

Virgin Media O2 (VMO2) has been on a bit of a customer service drive over the past year, but any fears that the UK telco would be switching out its customer services for AI agents should be put aside – or at least dampened – after it announced the creation of a new dedicated customer hub to handle repeat calls in a bid to boost first-time resolution for customers. The customer hub, which will be located at its office in Newcastle in the north-east of England, will comprise 150 agents dedicated to handling queries from repeat callers. It will offer tailored support in an effort to tackle complex issues through greater continuity and ownership. The announcement comes just a week after one of VMO2’s parent companies, Liberty Global, announced a partnership with Sierra to support the rollout of its conversational AI platform developer’s technology across Liberty Global’s operations in Europe – including the UK. Indeed, VMO2 made its own announcement explaining it would use Sierra to support calls from customers experiencing faults with their broadband service in a move that would free up human agents to focus on more complex cases. Only hours after that announcement, UK regulator Ofcom published new research that shows while people and small businesses are using AI to navigate phone, broadband and pay-TV markets, they have mixed views about it. This all follows an announcement back in February 2025 in which VMO2 unveiled plans to turn the company’s customer services around. Since then, it claims to have reduced complaints by more than 50% year on year and lowered the number of call transfers by 1.3 million in 2025, saving customers more than 400,000 hours of time chatting to VMO2 staff.

Veon has signed a 10-year agreement with Amdocs that will see the vendor modernise and manage Veon’s core technology operations as part of the operator’s strategy to reposition itself as a digital services operator across Eurasia. Amdocs will operate alongside QazCode, which is Veon’s software and AI subsidiary, to continuously modernise Veon’s operations with the aim of providing long-term cost predictability for the firm. The vendor will deploy its recently launched agentic operating system, aOS, to transform Veon’s operations, customer experiences and digital services, which already account for around 27% of Veon Group’s revenues. Veon CEO Kaan Terzioglu said: “QazCode has helped us build the technology and talent to bring better digital experiences to the millions of people who rely on us every day, and adopting Amdocs’ aOS capabilities is a natural next step in that journey. Our strategic partnership with Amdocs creates a strong foundation to keep improving people’s lives through digital and AI innovation over the next decade.”

Dublin, Ireland-based Beyond Now, one of the companies that features in TelecomTV’s most recent free-to-download report, The AI-Native Evolution of OSS/BSS, has launched its Agentic Revenue & Ecosystem Orchestration Platform, which the vendor says “connects capabilities traditionally fragmented across siloed IT and BSS environments and makes them available for people, systems and AI agents to act on through one commercial foundation.” The headless architecture platform boasts a shared catalogue, single data model and centralised commercial rules that, according to the company, “connect customers, products, partners, pricing, orders and services, giving people and agents the context to turn customer intent into personalised propositions, execute across the revenue lifecycle and increasingly transact agent to agent”. The platform is commercially available and is already being used by Hrvatski Telekom, part of the Deutsche Telekom empire.

Broadband network operators are facing some familiar challenges as they seek to use AI to automate their operations, according to the findings of a new report from the Broadband Forum. The industry body’s research “highlights poor device compatibility and inconsistent data as key challenges that must be addressed before broadband networks can become autonomous,” it noted. “The report reveals strong commitment to autonomous networking, with 94% of service providers planning to use software-defined networking (SDN) as part of their autonomous networking journey and 92% describing AI/ML integration as critical or important to these strategies. However, poor and inconsistent data from multiple systems was identified by 83% of service providers as a major barrier to deploying AI successfully in their networks,” stated the forum.

There’s a lot of excitement about the potential of smart (AI-enabled) glasses right now but also many privacy concerns, and these have resulted in the Norwegian government proposing a temporary ban on their use in some places, reports the Associated Press (AP). According to the AP, the Norwegian authorities have identified a number of places where “privacy is particularly important,” such as “healthcare facilities, or fitness centres with changing rooms and showers. A major focus is spaces where children congregate, such as schools, kindergartens, playgrounds and youth clubs,” it added. The news comes days after California’s governor, Gavin Newsom, vetoed legislation that would have penalised Californians for using smart glasses to record people without their permission in changing rooms, doctor’s offices and other spaces people generally consider private, stating in his explanation that the proposed bill’s definition of a wearable recording device was too broad.

- The staff, TelecomTV