What’s up with… Kyivstar, Taara, BT

• Veon’s Kyivstar heads for the Nasdaq
• Google-developed wireless tech firm goes solo
• Reports suggest Bharti’s Mittal might seek to increase his BT stake

Kyivstar, the Ukrainian telco that is part of the Veon Group, is to become a publicly traded company on the US Nasdaq stock exchange with a valuation of about $2.2bn as the result of  a merger with Cohen Circle Acquisition, a special purpose acquisition company (SPAC). Once the merger is completed (which is expected to happen in the third quarter), Kyivstar will be listed on Nasdaq with the ticker symbol KYIV. Veon will retain a majority equity stake, around 87%, in the Ukrainian operator, which has some 23 million mobile customers (48% market share) and 1.1 million fixed broadband customers (14% share), and generated annual revenues of $919m and adjusted EBITDA of $515m in 2024. “Kyivstar’s listing on Nasdaq will be more than a financial milestone – it will mark a defining moment for Ukraine’s economic future,” noted Veon chairman and founder Augie K Fabela II. “As one of Ukraine’s largest private sector investors, Veon is proud to pioneer this historic step, offering US and global investors a unique opportunity to support and participate in Ukraine’s economic growth and resilience,” he added. Veon Group’s CEO, Kaan Terzioglu, added: “As part of Veon, Kyivstar has a strong track record of consistent performance, turning challenges into growth opportunities for Ukraine, for its customers, partners and team members. Investing in Kyivstar now becomes more directly accessible to international investors as Kyivstar progresses towards becoming a Nasdaq-listed company. We believe that Kyivstar’s strong financial profile, visionary strategy and robust governance structure, developed over many years as a Veon Group company, will be appealing to international investors.” Kyivstar CEO Oleksandr Komarov stated: “Through the war, the Kyivstar team has succeeded in spearheading the resilience and rebuilding of Ukraine’s infrastructure while building the foundations of its future success as a leading digital services company in Ukraine. The progress towards Kyivstar’s listing will allow us to share this growth story with global investors. With our highly experienced and motivated team, we look forward to completing the business combination with Cohen Circle to embark on the next phase of Kyivstar’s growth.” SPAC deals (an alternative to a full IPO process) were all the rage a few years ago – there were more than 610 in 2021 – but the number of subsequent valuation crashes and business failures mean SPAC deals are now few and far between (with just 57 in 2024).   

High-speed wireless technology developer Taara has attracted investment (of unspecified value) from external investors in a funding round led by Series X Capital and is spinning out of X (aka The Moonshot Factory), the R&D unit of Google’s patent company Alphabet (which will retain a minority stake). Taara’s technology, dubbed Lightbridge, uses point-to-point wireless data links to transmit information at up to 20 GBit/s over distances up to 20km and was the main element behind the now-defunct Project Loon, Alphabet’s experimental dirigible network development – see Alphabet brings Loon’s optical wireless tech to terra firma. According to a blog written by Taara’s CEO, Mahesh Krishnaswamy, Lightbridge “brings fast, fibre-like internet access to areas where it’s too difficult or expensive to install traditional fibre, like in dense city neighborhoods, over rivers and seas, or across rugged terrains and national parks. Taara’s Lightbridge units… require only a few hours to set up, without the time and cost associated with digging trenches or stringing cables.” The technology is already being used by India’s Bharti Airtel, African operator Liquid Intelligent Technologies and Liberty Networks and is being checked out by Vodafone Group and T-Mobile US. News that Taara is becoming an independent company follows its unveiling of a silicon photonic chip during the recent MWC25 show in Barcelona. 

The Financial Times (FT) (subscription required) has reported that Indian billionaire Sunil Bharti Mittal is keen on increasing the 24.5% stake in UK national telco BT Group that his Bharti Enterprises conglomerate already holds. In August 2024, Mittal’s Bharti Global, the international investment arm of Bharti Enterprises, agreed to acquire the near quarter stake in BT held, at the time, by Patrick Drahi’s telecom vehicle Altice in two tranches – initially 10% with a further 14.5% added later. At that time Drahi was under severe pressure to sell assets and raise cash as Altice group debts ballooned to more than €60bn. Drahi divested the BT shares after steadily building up his stake in the company over a three-year period: We’ll likely never know what the Altice owner’s plans were for his stake in the UK telco, but we do know that the BT board was worried that it might face an aggressive takeover bid. Bharti Enterprises is now the operator’s largest shareholder and indications are that Mittal will buy more stock as BT continues to cut costs and reorganise under the leadership of CEO Allison Kirkby. Mittal previously indicated that he would not consider buying more shares in BT until the complex process associated with acquiring Altice’s stake in BT was completed, but it seems that day is now getting closer. Were Mittal to choose to increase his shareholding in BT to more than 30%, he would be legally bound to make an offer to acquire all of the BT stock and take control of the telco. Kirkby has long maintained that Bharti will not be a “passive shareholder”, noting that Mittal “definitely will have a lot to add considering the business that he has built in India and internationally”. In India, Bharti Enterprises is the majority stakeholder in Bharti Telecom which, in turn, is the largest shareholder in India’s second-biggest telco, Bharti Airtel. A company spokesperson told the FT that Mittal “currently has no plans to buy any further stake above the 24.5%”. BT, meanwhile, is keeping schtum. Watch this space.

Companies in Europe are developing new technology that will allow subsea network operators to monitor and protect fibre cables by “listening” to them and pinpointing anomalies, such as sabotage. The technology is based on measurements of changes in the acoustic energy travelling through fibre when the signal is disturbed. When light pulses move through a fibre strand, very small ‘echos’ of the light bounce back along the line. When monitored, the disturbances appear on screen as visible blips, longer lines or starlike bursts. A BBC article explains that the new measurement system enables its users to gauge the location and approximate size of a ship or other vessel passing over a submarine cable, together with its direction of travel. That data can then be correlated with satellite imagery and automatic maritime identification system (AIS) records, which most ships broadcast continually. Those that don’t match are automatically suspect. The ‘sound’ of an anchor being dropped is also recorded and its trace is particularly evident. However, the efficacy of fibre-optic sensing is limited and localised, which means cables need signal-listening devices (called “interrogators”) to be installed within them every 100 kilometres (km) or less along an entire cable. Paul Heiden, the CEO of Optics11, an Amsterdam-based  company specialising in fibre-optic acoustic sensing systems, says ‘listening cables’ installed just to monitor marine (and submarine) activity, perhaps at a 100km distance from ports, or in the vicinity of telecoms cable and gas pipelines, would be a cheaper alternative. Acoustic sensing is becoming more important as incidents of mysterious damage to cables increase, not least because prevention is better than cure. Fibre cables already have some protection from outside interference and ‘armoury wire’ can be deployed to help protect strategic assets but they are more expensive and by no means invulnerable to accidental damage or deliberate attack. 

Mobile World Congress is an incredibly important event in the telecom calendar but for those who attend the show, it can make it much easier to miss interesting and important industry developments and this editor, for one, is still catching up with recent developments, such as the rebranding of Canadian firm Sandvine as AppLogic Networks. These days the company presents itself as an over-the-top application classification and quality of experience solutions vendor, but when I first started engaging with the company it was known as a deep packet inspection company (and that’s still what its technology does – analyses and reports on data traffic to provide useful insight to network operators). But these days, deep packet inspection (DPI) is a toxic term for many reasons but mainly because some have put DPI technology to nefarious uses. In fact that’s what ultimately led Sandvine to be placed, in February 2024, on the US Bureau of Industry and Security (BIS) ‘entity list’ of companies that are barred from using US technology. Once it got over the shock, the company decided to revamp, withdraw from multiple markets, pledge to only do business in “democratic jurisdictions” and introduce multiple levels of corporate transparency, moves that led to its removal from the entity list in October last year. By that time the company had a revamped strategy and new investors lined up and in November it entered into a restructuring process, which it completed in February this year. Now it has a new name, a new CEO in Mark Driedger (who was COO for more than nine years prior), and has just released its latest annual Global Internet Phenomena Report (GIPR), which provides “in-depth insights into global internet traffic patterns and the applications driving that traffic”.  

Who’s a clever boy then? A UK neuroscientist claims that AI chatbots can think and are not just “stochastic parrots” that have been trained to seem intelligent but are actually as thick as a brick. In a new book, These Strange New Minds: How AI Learned to Talk and What it Means (reviewed here by the Washington Post), Christopher Summerfield, professor of cognitive neuroscience at Oxford University’s Wadham College, claims that while large language models (LLMs) are not sentient (yet), they do exhibit characteristics similar to human brains. They are error-prone and make predictions based on little empirical evidence and leap to imperfect conclusions that can often be wildly incorrect. Sounds just like an average day in the TelecomTV editorial suite... However, AI chatbots can, and do, hold forth on almost any subject and partner with humans in seemingly coherent conversation. This, Summerfield believes, shows that AI chatbots demonstrate an ‘understanding’ not only of what words can be strung together to make some sort of sense but also what formulated thoughts (ie ideas) belong together. He claims the structure of language maps onto the structure of reality as humans perceive it, and so predicting what to do in a massive range of possible scenarios requires a model of not just how language works but also of how the world works. The professor reckons the correspondence between language and reality can explain what AI researchers call “emergence”, which is the ability of LLMs to provide answers to questions and solve problems, the solutions to which are not in their training data. They shouldn’t be able to do it, but they can. That said, many of the answers provided are “hallucinations” – that is, spurious and totally fictional gibberish. Summerfield says these flights of fancy should be called ‘confabulations’ – memory errors comprising fabricated, distorted or misinterpreted memories. They differ from lying in that the person – or an AI chatbot – is not consciously attempting to deceive. Well, it couldn’t, could it? It isn’t conscious, it’s not very bright and is a work in progress.

– The staff, TelecomTV