At FutureNet World, Rakuten and Telus pushed for bold AI-led reinvention, while BT and Orange argued for a measured platform-driven evolution. But beneath the contrast, a shared strategy is taking shape – programmable networks, available in flexible platforms, plus proper teamwork to monetise AI in vertical solutions. Don’t just sell connectivity, they all said.
In sum – what to know:
Different instincts – Rakuten and Telus promoted more radical reinvention at FutureNet World in London; BT and Orange presented strategies for pragmatic evolution.
Common models – but telcos, generally, are converging on a common model: API-driven platforms layered on core network assets, geared to support AI services.
Existential changes – unlike cloud or edge, AI is viewed as a non-optional, board-level priority; but time is short, and the chance to capture value is urgent.
There was some good ying-yang insight about different telco approaches to AI on a morning panel session (‘can telcos monetise the AI gold rush?’) at FutureNet World in London yesterday (April 22), with Rakuten and Telus on one side, making the case for sweeping technological and organisational change, and BT and Orange on the other, presenting more pragmatic platform strategies. Maybe in the end, they were all saying the same thing – build flexible platforms on critical assets, and new partnerships alongside to make a virtue of connectivity in vertical solutions. Maybe it was just cultural; but they claimed variously to be in agreement and disagreement, and their opposite instincts seemed to say something about their respective legacies, attitudes, strategies, geographies.
Per its reputation and position, Japan-based Rakuten was the cat among the pigeons, here – represented on the panel by Faiq Khan, head of EMEA sales at Rakuten Symphony, the firm’s platform vendor business, selling the Rakuten Mobile blueprint in international markets. “There’s no doubt telcos are perceived as trustworthy partners for enterprises… [But] we often confuse customer trust, customer retention, and customer loyalty. I use an operator in the UAE; I use it for another service in the office. I trust it very well; it does everything to retain my business. But is it able to get more wallet-share from me? That is the question with AI: are telcos, for once, going to get a bigger wallet-share? [Because] unless and until they have monetizable solutions, they won’t make more money.”
Rakuten is a “rule breaker”, suggested Roy Chua, founder at analyst house AvidThink, and chair of the session. It also has a “reverse play” in the telco space insofar as it is an e-commerce and internet services conglomerate (like Japan’s Amazon) with a mobile business on the side. “In some sense it’s not fair,” said Chua. “Yeah, it’s not fair,” agreed Khan. “And it is absolutely easier for us, on a smaller scale, to make it happen. But it sets the tone – whether you’re an Orange or BT of the world. Because you want to get there, too. [But it’s like with] smart fridges – very good technology, wow, but how much monetization has happened? This industry is great at assuming that its technology is wow, but how it monetizes the tech has been a big question… There is an opportunity, no doubt.”
Leadership issue
When asked by Chua – going around the horn, to close the session – what progress telcos will have made with AI in 18 months, Khan was punchier; the AI opportunity might even have gone by then, he suggested – unless telcos shake their traditional sloth. “I’ll give a slightly twisted answer,” he said. “We have an opportunity to sell AI – or, even, to be an AI company that happens to have a network. This is where I have a bit of disconnect [with the] panel… I listen and respect and appreciate that some big telcos come with massive legacy infrastructure, and it’s not easy to change overnight. But the mindset change is still slow… There is a generational issue – where Generation X/Y, whoever is [in charge], has a lot of bias from previous generations and remains very tactful and careful.”
Japan, described as a “very careful environment”, has been “very bold”, he said – whether as a country-economy with AI, or as a country represented by a company with AI. Between them, Rakuten Mobile and Rakuten Symphony made $250 million of incremental revenue from their “AI environments” in 2025, he said. Every other telco should copy (and buy the platform from Rakuten Symphony, presumably), as well as it can – is the implication. The wider industry has to take a “generational leap in decision making and fearlessness to achieve real monetization”, said Khan. Indeed, there was a subtext here that this opportunity should not be taken for granted; it is a one-off, and “existential”, said Hesham Fahmy, chief information officer at Canadian operator Telus.
“Edge and cloud were optional technologies. There was value, but you didn’t have to [adopt / sell them]; it wasn’t existential. But tell me an organization for which AI is not an imperative. It is a board level discussion. This is happening, right? It is existential – for everybody. It’s like connectivity was 30 years ago. Once it is there, everybody needs it. Same thing here… The market is captive. It is not optional… Everybody has to figure out how to get in the AI game, and use it.” And there was agreement, too, in the red corner that telcos’ go-to-market strategies with enterprises should ‘go beyond connectivity’ – per the FutureNet theme. “Your go-to-market motion can’t just be to sell connectivity for AI. If you sell that, it will fall flat,” said Fahmy.
He went on: “It has to be a part of a vertical solution. And whether we build our own, which is probably not a good idea, or partner with others, the go-to-market is to be part of an end-to-end solution – where you enable that solution. And then you’re going to get the pull-through, right?” Right, agreed Khan – and BT and Orange, as well, actually. Khan said: “Are you looking to sell a service to a logistics company [or just] connectivity? [Are you selling] timely delivery as a service – so they get a connectivity package, but [also] efficiency [guarantees] and edge inference? That is the opportunity… We should not be selling IT networks to retailers. We should be selling loss-prevention AI – about how their camera analytics and edge inference bring end-to-end alerts for both the retailer and the buyer.”
Platform approaches
This is where the conversation was best – and it was a genuinely good discussion. Khan, here, was responding to a statement from Orange, made in response to a statement from BT. Both were explaining their company positions – to build a programmable AI-threaded software platform on top of their core AI-geared infrastructure, to expose network capabilities and climb the “value chain” with enterprises. Both are worth replaying, if only just to set the battle lines. Here’s Sef Tuma, chief technology and information officer at BT International: “A lot of telcos try to build the whole stack, sell the whole stack – and their margin erodes because of what they’re buying. That’s not a future model, at least for us. [We will] make money from the use of our platform – to have a large-scale volume of bandwidth for AI.”
He went on: “My job is to make sure it is optimal for every enterprise, however they want to work – whether to automate a factory, or [run] low-level finance systems. [Enterprises have] different requirements… I’m not going to build [their] large-scale transformational programmes… Because that is a significant investment in verticals… or skills and capabilities that aren’t network-based. Partnering is an approach I care about, but… not to sell my network. I don’t think that works. What works is to build a consolidated solution that no one else can to optimize the enterprise [by] optimizing the network for AI – in terms of sovereignty [and whatever is] critical to them. That’s a page from AWS and Azure, which don’t build that many industry solutions, but definitely help [enterprises to] make their discs spin.”
So as above, BT’s proposition for multinational enterprises is an API-driven network (’global fabric’) platform that partners can build AI services on – and to make money from enabling developers and enterprises, and not just selling airtime. “We will be the foundation for those [solutions] as they use AI, and even as they use our AI engines to optimize their network to do things. That is a way to scale new monetization… We will work with any managed service provider [to] be critical to the outcomes they provide,” said Sef. Is this different from the beyond-connectivity model, actually? The language is less radical, but the end goal seems the same. But, even here, even in the blue corner, even amid the bonhomie in London, there were differences. Orange wanted to “agree and disagree”, it said.
Miguel Alvarez, chief data and AI officer at Orange Business, responded: “The platform approach is essential, and something telcos need to learn from the cloud…. [But] different market segments want different things. Companies don’t want everything from the same provider, but they do want it to be simple. They don’t want a super technical product… That’s why hyperscalers partner with local players to bring the cloud to [enterprises]. AI will be the same… The mistake, traditionally, was we built super-standard solutions one way, and everything [else] was super-custom… So we had these overblown packages that people wouldn’t pay for. With a platform, you can [provide] professional and managed services to deliver simplicity… and work with partners to provide [custom] services indirectly.”
This article is continued here.
