Brice Evin, CFO at FLAG, the privately owned global subsea cable system, talks about new financial and operating models in the face of geopolitical tensions, AI and next-gen of hyperscalers
Brice Evin is CFO at FLAG – one of the world’s largest privately-owned subsea cable systems. FLAG’s origins stretch back to the 1990s with a link between the UK and Japan. Back then it was called Fibre-optic Link Around the Globe (FLAG) and privately funded undersea cable systems were rare. FLAG launched commercial services in November 1997.
Evin says the company has undergone tremendous changes in its 40-year history and “the new life of flag started in 2020 after a restructuring period, new management coming in, and so on”. In August the firm promote Paul Abfalter to CEO, with effect from 1 October 2026. Currently Abfalter is FLAG’s Chief Strategy & Revenue Officer and will succeed outgoing CEO Carl Grivner who become Non-Executive Chairman.
FLAG serves three kinds of customers: hyperscalers; telcos (like Vodafone), content and new media service providers; and corporations. It has operations in more than 180 countries with infrastructure stretching from the US’ East Coast to Europe, on to the Persian (Arabian) Gulf with links to Asia, and from Asia to the US’ West Coast.
In his long career in telecoms, Evin has had stints at Vodafone and Alcatel Lucent before becoming CFO of Global Cloud Xchange (GCX) in August 2023. He still holds this position: on 1 April 2025, GCX rebranded as FLAG and GCX became the brand for FLAG’s Managed Services division. Throughout growth of data traffic has accelerated that’s still the case now, Evin says, driven by AI.
The partnership model
A major change in the subsea cable industry now is the move towards partnerships to build subsea infrastructure. Evin notes, “Way back, FLAG owned seven subsea cables – Hawk, Falcon etc – built at a time when…building subsea cables was seen as very innovative and there was Incredible demand.
“Nowadays the industry is much more fragmented. I’d say, three factors are driving the shift to partnership. The first is growing geopolitical tensions. Number two is regulation; many countries are reinforcing mobile or fixed telecoms, and submarine cables for long distance. Fibre regulation is tightening, and countries are getting more and more protective, including from a sovereignty perspective.”
The third part, according to Evin, is that hyperscalers have been building their own infrastructure but now the likes of Google, Meta and others are rethinking strategy because some infrastructure is in increasingly dangerous places which threatens reliability.
Accidental damage and sabotage
For many years the primary cause of cable damage was shipping and their anchors, plus natural causes such as undersea landslides. Now there is outright sabotage on an unprecedented scale and war in the Middle East which is the world’s busiest undersea corridor. Regardless what causes disruption, resilience that depends on a proprietary portfolio of cables is no longer a viable strategy, Evin says, hence, “The industry is moving to an ecosystem of partnership so customers have access to multiple paths.
“Maybe 10 years ago, they just wanted one path from Europe to Asia; now they need three, four or five for resilience via a diverse path end-to-end. They see resilience, especially in this AI age, as critical.”
Evin says network operators want to partner with “a multitude of local and regional players, and that’s the strength of FLAG. We offer very good local anchoring, ties with the authorities…reach between regions because we’ve been operating for a long time.
“We’re developing a partnership platform where we can deploy up to 30 passes between Europe and Asia, for instance. This is really the big players moving from running their own traffic on one cable they own, to trying to divide as much traffic as they can between different parts.”
Control through partnerships
Does this mean FLAG is becoming a kind of aggregator and orchestrator of capacity rather than an owner of subsea routes?
“No”, Evin says. “We don’t want to rely on partners to deliver quality of service. We want to own and control through the partnership, we’re building and deploying our own assets. Instead of building one cable for maybe $500 million that would take probably between five and 10 years from initiation to ready for service, we are investing incrementally with our partners and through our partners, while retaining end-to-end control of the network.”
He continues, “That’s the key piece: if an operator has to interact with so many parties, they dilute responsibility and network control across many parties. We’re keeping and retaining control of the network so they know we’re a one-stop shop, providing the service they need on a multi-path solution with one throat to choke.”
Anchor tenancies are key
Another goal of the multi-partnership versus ownership is anchor tenancy: “For example, if we want to serve a customer wanting a route through Iraq and Turkey (we call this the northern route from India to Europe), we would talk in advance to hyperscalers and new media companies and they would finance half the capacity on innovative routes,” he explains.
“It’s very important for us before we sign to have the backing of a consortium of seven banks, say, and they ask us to de-risk the investment. Partnership goes well with the anchor tenancy. We are seeing an accelerating demand for parties to become anchors.”
Evin continues, “Anchor tenancy works best with the end customer, probably arising from the fact that the demand for data is urgent for the end customer. Someone like Google or Meta has a business to attend to, so they can’t wait. They have forecasts of data demand for the next six to 12 months and absolutely need to have a solution in place.”
Talking telco
This is different from the telco world which Evin sees as “a bit more mature…If we’re working with Vodafone and other telcos, their domestic activity is well established in terms of data bandwidth and demand. There is definitely more activity coming from the end customers whose data demand is generated by their own customers internally.
“Advertising is key for someone like Meta and Google – they rely on ads in real time and so cannot compromise on data latency. For them, it’s critical to have diversity and they’re more aggressive about accessing new solutions through partnerships.”
AI changing traffic patterns
“As we [re]started last year, we refreshed our entire strategy for FLAG, which we call vision 2030. It was really about where the new demand is going to be,” says Evin. If you project five to 10 years forward, the design of today’s data traffic is that AI needs to run large language models which are largely based in the US.”
At the moment, regardless of where you are in the world, if you engage with Gemini or ChatGPT, your request is very likely to go to the US then come back to you. FLAG’s strategy is putting edge data centres next to connectivity points. Based on studies, we believe the workload currently in the US will spread out across the world. The Gulf region will be a very important concentration of future AI models because it has the lowest latency points if you want to go from Asia to Europe, or south to Africa,” he says.
Middle East corridors
Evin continues, “Without naming anyone, the big OTTs are asking us to double down on this corridor through the [Arabian] Gulf. Whether that’s the northern route – north of the traditional routes from the Med[terranean] down to the Red Sea, around the Persian Peninsula onto India. Or a mid-route that would go to Oman and Saudi, and cross the Saudi Peninsula east-west and end up in the Red Sea, avoiding Yemen and those pinch points via a terrestrial path.
“We have lots of partners in Saudi and Iraq, Jordan and Israel in were actively building the next generation of connectivity and highways that will bring a diversity. This region and this corridor is bang in the centre between Europe and Asia and will remain very central, very important to the future of connectivity between across Asia and Europe.”
Evin acknowledges India has not developed as fast as some other Asian economies but it is very strong in terms of IT, engineering, workforce and big corporations developing their own audience. He adds, “Also, 50% of the traffic from India goes west and 50% goes east, so it’s very important to connect India in both directions, to Europe and the US respectively. India the most populous country on Earth and developing at a very high pace. We see that traffic and the demand for connectivity and data centres is huge.”
Evin elaborates, “If we look at Giga watts per head in India, it’s a lot lower multiples than other regions. They have plans to erect new data centres [on a massive scale] and usually data centres require connectivity – you need to build the highways to connect to them, otherwise they are just a warehouse.”
Southeast Asia and new routes
FLAG also thinks the corridor between Southeast Asia and the US, across the Pacific from the US to Singapore, will become more important. In June, FLAG announced the launch of a new route “set to reshape connectivity between South Asia and Southeast Asia”.
In future, from southeast Asia, FLAG’s infrastructure will reach into northern Asia, beyond the developed markets of Japan and South Korea. Evin says, “Singapore is highly congested but Malaysia, for example, is more and more attractive and the appeal of the Philippines and Indonesia will also increase.” Indonesia is the world’s fourth most populous country and the size of Europe in terms of landmass. There’s going to be a lot of demand from this area going to Los Angeles and the US’ West Coast.”
The data centre business
The growth of data centres, largely driven by the rise of AI, is central to FLAG’s strategy. In the past, the trend has been to centralise data centres, Evin comments, “Now we see more and more content needs to be streamed or stored as close to customers as possible, so connectivity deployment changes as well.
“We get more demand from customers, including larger cities, to have smaller, more scalable data centres at the edge point of our networks. We’ve worked on many projects to build smaller size data centres located close to our landing stations – and we land in more than 40 countries. Some places like the UK we own the land, other places we lease it.
Building data centres is a big shift in FLAG’s operational and business models? “Yes and no,” he says “because our strategy has been on connectivity, subsea cables, then a mix of subsea and terrestrial. At the same time, we’ve always operated big, complicated landing stations, which are similar to data centres in that you have power, cooling, connectivity, the skills and so on.
“The only difference is, instead of operating the subsea cable ourselves, we put the customers’ servers there to operate whatever business they’re in. We already have several locations in operation for some OTTs. Now we are expanding, offer customers data centres ‘bundled’ with our connectivity in countries where the OTTs don’t want to build. For example, we’ve been on the ground in Iraq, Saudi, Jordan, Oman and Dubai, Bahrain and Qatar for decades. They’re very happy for us to propose solution that combine connectivity and data centres to store their workloads and content, but also in future, their AI workloads.”
Next-gen customers in the next decade
How do you think the picture will change between now and 2030 or 2025? Will there be more big changes or that we will see a gradual shift to the edge and continuing diversity of routes?,
“We are preparing ourselves for the next generation of OTTs,” states Evin. “Today we are familiar with the Magnificent Seven. Some will continue to be super successful, but some will lose some of their shine or drive. We are already looking at the next generation of companies that were born in this AI bubble and will become the hyperscale companies in future.
“There will be a shift in terms of our customer landscape. We are already starting to see some tier two and tier three companies emerging. Mostly they are focused on content security, moving away from the typical distribution of content.”
He continues, “We’re going to see a new type of content distribution, a new generation of media players and who are very conscious of data sovereignty and security – this is becoming a theme. We also believe that OTTs will move further away from owning their own infrastructure – we’re talking with the big OTTs that are still rolling out massively but at some point they won’t have unlimited funding. Then they will be looking for new ways [to operate] and they will refocus, like the telcos have done, by monetising their radio towers and the network.”
FLAG predicts this will also happen to the big content companies. “They won’t be able to splash billions to roll out their own fibre,” Evin says. “That’s where we come in, because we are getting ready for acquisition and to lease capacity back from them. There will be a change in ownership, and that’s where the partnership model is coming into play. They will want to focus on content but at the moment the location of content is very US-centric and will need infrastructure for distribution.”
Satellite features in FLAG’s future
Evin says, “We have a big satellite company as our customer already because satellite is very powerful for providing local access: where you don’t have Vodafone coverage, you are happy to have satellite communication as the nearest point you can connect to. But when it becomes too long a distance, it doesn’t really work. High capacity communication from satellite to satellite doesn’t exist today. Vodafone itself is working closely with satellite as well.”
“Long distance satellite-to-satellite communication is a long time away, so I think we don’t see satellite in the mix in the midterm,” he says. By 2030, I don’t see satellite as competition, I think they are really buying capacity. From Europe, if you want to use SpaceX to call you the US, there’s a 99.9% chance the link will go subsea – so Europe-to-Europe satellite link, European-to-US subsea, then back out onto a US satellite link again. In the near term it will remain an access communication system, not a long-distance communication system.”
Evin concludes, “Ultimately, as we get closer to ubiquitous communication, we won’t care what technology our communications run over so long as it’s affordable, reliable and of good enough quality.”
