• Deutsche Telekom highlights 6G-TakeOff advances
• European operators boast sovereign service developments

In today’s industry news roundup: After three years of R&D, the DT-led 6G-TakeOff project unveils its outcomes; the enterprise services arm of Proximus teams up with France’s Mistral AI for European sovereign digital service development; Bouygues Telecom deploys a sovereign agentic AI platform; Veon’s software unit teams up with a specialist partner to develop local language LLMs for its various operating markets; and more!

The Deutsche Telekom (DT)-led 6G-TakeOff project, which was first announced in October 2022, has presented the results of three years of R&D efforts that have been focused on “3D networks and their key components”. These three-dimensional networks comprise base stations located on the ground, on high-altitude platform systems (HAPS) and aboard satellites. “The project focused on the holistic view of a 3D network and the question of how the various network elements can be connected to each other in a unified 6G architecture. By combining and intelligently coordinating the various access technologies, optimal access to connectivity is thus enabled for every application. The results of the project are an important part of basic research for so-called non-terrestrial networks (NTN) and will be incorporated into the standardisation of the future generation of mobile communications,” noted DT in this press release. The R&D work was undertaken by a consortium of 19 partners, including telcos – DT and one of its domestic rivals, O2 Telefónica (Telefónica Germany) – technology and networking companies – including Airbus Defence and Space GmbH, NXP Semiconductors, Rohde & Schwarz and Boldyn Networks – and multiple research institutes and universities. The partners developed several demonstrators to test the feasibility of different solutions, including: Device handover in the 3D network; the deployment of mobile edge computing (MEC) services (including on NTN platforms); and feederlink technology for ground stations and unmanned aerial vehicles (UAVs), or drones, which require connectivity to a core network platform via high-speed “feederlinks” to ground stations. For the feederlink solution, beamforming antennas were developed and tested. In addition to the demonstrators, the project achieved seven patent applications, highlighting “the innovation potential of the consortium”, noted DT.  Thomas Lips, senior VP of RAN disaggregation and enablement at Deutsche Telekom, stated: “The 6G-TakeOff project has helped us better understand the practical challenges of integrating terrestrial and non-terrestrial components into a unified 3D communication framework. It offers valuable insights on how future 6G systems could improve service continuity, resilience and capacity wherever needed. The project has laid a strong foundation for further cross-industry cooperation towards 6G.”  

Proximus NXT, the enterprise services division of Belgium’s national operator Proximus, has teamed up with major French AI developer Mistral AI to “launch a comprehensive AI offering targeting not only the Benelux region but also other European countries. This partnership gives organisations access to solutions, such as AI Studio, Le Chat and Mistral Code, developed by Mistral AI, while allowing them to leverage Proximus NXT’s AI expertise and secure infrastructure,” the operator announced. “Hosted by Proximus NXT, this turnkey solution is designed for organisations that want to benefit from the power of Mistral AI’s suite of artificial intelligence products while relying on a sovereign infrastructure. The aim of this offering remains unchanged: To make AI ‘made in & for Europe’ accessible – high-performing, scalable, secure and truly sovereign.” Fabrice De Windt, Proximus NXT’s interim enterprise IT services and segments lead, stated: “Artificial intelligence is a powerful driver of competitiveness, but it must remain controlled, safe and accessible. Together with Mistral AI, we offer organisations a high-performance, sovereign solution that can adapt to their needs, whether they operate in the Benelux or in other European countries.”

Following a successful trial, Bouygues Telecom recently struck a deal with Toulouse, France-based Prisme.ai to help the French telco deploy a unified “secure agentic AI platform… that structures and industrialises AI usage for Bouygues Telecom employees,” the operator noted in this announcement (in French). The deployment of the AI Studio, as the system is called, is “a true cornerstone of the operator’s digital strategy” as it “allows business teams to create and manage their own AI agents… The goal is to move beyond conversational AI to enable teams to design, manage and monitor their own AI agents that meet the specific needs of each business unit.” The AI ​​Studio was launched in mid-September and is now fully operational and being increasingly adopted throughout the company as training programmes are completed. It has more than 5,000 registered employees, with an average of approximately 830 unique users per day. According to Bouygues Telecom, in just a few weeks, more than 800 AI agents were created by business teams, including 51 agents deployed at scale and accessible to everyone via the AI ​​Studio catalogue. “Usage is strong and accelerating rapidly, with over 10,000 questions and answers generated each week, illustrating the rapid and natural adoption of AI agents by employees,” noted the operator. Stéphane Allaire, director of innovation and head of the generative AI programme at Bouygues Telecom, stated: “With Prisme.ai, we have a platform capable of protecting the data used by our AI agents, integrating with all our systems, and deploying business agents on a large scale. This partnership allows us to innovate quickly but with solid governance and total control over our usage.” Wafaâ Amal, CEO of Prisme.ai, added: “This partnership illustrates a turning point in how large organisations are adopting artificial intelligence. With Bouygues Telecom, we are demonstrating that widespread adoption of AI is only possible when business teams can create their own agents, within a framework controlled by the technical teams… With Bouygues Telecom, we are laying the foundations for agentic AI that serves performance, transparency and European sovereignty.”

QazCode, the software development unit of Veon, the digital service provider with operations across Europe and Central and South Asia, has signed a strategic cooperation agreement with “sovereign AI firm MeetKai to develop and train next-generation large language models that will enable deployment of local-language agentic AI services across Veon’s operating markets,” the operator has announced.“This strategic sovereign partnership is designed to accelerate the creation of nationally governed AI capabilities that are culturally fluent, linguistically inclusive and securely operated within each country’s borders. MeetKai’s sovereign AI platform provides an end-to-end, customisable national AI stack – enabling countries to govern how models are trained, hosted and deployed, aligned to domestic regulatory and data sovereignty requirements,” noted Veon. The service provider says the collaboration will help it to develop “safe, locally relevant AI digital services for over 150 million customers across Kazakhstan, Uzbekistan, Ukraine, Pakistan and Bangladesh,” and is “expected to pave the way for a new generation of AI-based digital services” developed for specific industry verticals across Veon’s markets. “AI is a critically important and pervasive technology, yet today most global resources are concentrated on a small number of widely spoken languages,” stated Oleksii Sharavar, CEO of QazCode. “This partnership with MeetKai allows us to build sovereign, locally governed AI foundations that will help catalyse future AI developments across our markets.” James Kaplan, co-founder and CEO of MeetKai, added: “True sovereign AI means countries decide how their models are built, governed and deployed. Through this partnership with QazCode and Veon, we’re proud to serve over 150 million people with national AI stacks that respect data sovereignty and deliver real, committed value in their language,” added Kaplan.

– The staff, TelecomTV