- Trust, sovereignty and execution are the name of game at FutureNet World in London today
- Telcos have largely accepted they won’t out‑compete hyperscalers on infrastructure or AI scale
- Vodafone doesn’t buy GPUs anymore; it rents them, according to Scott Petty, CTO at the operator
FUTURENET WORLD 2026, LONDON – For more than a decade, telecom executives have been promising digital transformation, reinvention and a brighter future just around the next architectural corner. Digital transformation gave way to cloud transformation, which gave way to AI transformation. Now, the industry is trying on new phrasing: evolving operating structures and shifting the business to thrive.
Don’t get me wrong — they are still under pressure from shareholders and customers. But during a panel at FutureNet World, moderated by Mark Newman, chief analyst at TM Forum, senior executives from Telefónica, Vodafone, Orange and Blue Planet offered candid insights about what telcos can realistically compete on — and what they cannot.
One big one: Scott Petty, CTO at Vodafone Group, noted that Vodafone doesn’t buy GPUs anymore. “We rent them because our workloads look a lot like IT workloads, and they lifecycle very, very quickly,” he said.
In other words, the arms race is over. Telcos are no longer pretending they want to own the entire AI stack.
On top of that revelation, a long‑running debate has effectively been settled: Telcos are not going to beat hyperscalers at economics.
“The go‑to‑market proposition is really ‘connectivity-plus,’” said Laurent Leboucher, EVP Networks and group CTO at Orange, when pressed on how telcos differentiate in an AI era. “Connectivity plus providing AI capabilities that can be done in a way which is considered secure enough by our customers.”
That’s a notable shift from past years, when edge compute and AI platforms were sometimes pitched to rival Amazon, Google or Microsoft.
Where is the leverage?
Where they do see leverage is in operational discipline — an area that has undermined profitability for years. AI, automation and autonomous networks are less about shiny new services and more about fixing the mess underneath.
“The forcing function on AI is actually a bigger catalyst that’s forcing a change everyone has wanted to make for years but hasn’t been able to,” said Joe Cumello, SVP and GM at Blue Planet. That change? Cleaning up decades of operational sprawl. “Collapsing 50, 70, 80 applications in OSS… You can’t apply AI against bad data.”
Cumello cited a customer cutting $14 million per month simply by cleaning up data and decommissioning systems — this is the kind of financial impact operators need.
Petty reinforced that view, arguing that the companies executing well on AI aren’t talking about future products. “They’re talking about how they’re harnessing AI and the value it’s delivering today,” he said. The opportunity, Petty added, is to “make us better businesses — to service our customers better, to run our networks better.”
Evolving operating structure vs. digital transformation
This is where the “evolving operating structures” language starts to feel more doable than digital transformation ever did.
Telefónica Group CTIO Andrea Folgueiras described how AI‑driven network planning has compressed timelines from months to hours. “Our aim should not be to optimize tickets,” she said. “We should try to avoid tickets.”
That mindset shift — fewer humans firefighting legacy complexity, more automation preventing problems outright — may end up being more consequential than any single AI product launch.
So, what does “beyond connectivity” actually mean now?
Partly, it’s about trust and sovereignty rather than latency, which was a hot topic ahead of Mobile World Congress 2026 in Barcelona.
The operators on stage here in London spoke cautiously about edge compute, acknowledging that ultra‑low latency is hard to monetize outside very specific use cases.
Sovereign reigns supreme
Now, the message, especially in Europe, is demand for sovereign, regulated and trusted environments — particularly for public sector and critical infrastructure customers.
“Sovereign AI without a sovereign network doesn’t make a lot of sense,” Petty said. “That’s a real opportunity for telecommunications providers — to provide certainty around how the network’s operated, where it’s operated and who can access the data.”
Security also remains one of the few areas showing consistent growth, with Orange highlighting network‑embedded security services such as DNS filtering and number verification, which is much less glamorous than AI factories, for example, but far closer to real buying behavior.
Realism is the name of the game, it seems. Capex budgets haven’t exploded. Petty noted that Vodafone is funding AI by decommissioning legacy platforms and reclaiming oxygen from decades of technical debt.
“We have to learn lessons from the past,” Petty said, “and never allow that technical debt to accrue again.”
Note: This story was updated on April 21, 2026, at 12:17 p.m. ET.