MWC: Can telcos turn digital sovereignty into revenues?

European telcos’ sovereign AI investments could falter if governments do not get behind them and start buying their services. That was the warning sounded by Timotheus Höttges, CEO of Deutsche Telekom, during a panel discussion of digital sovereignty at MWC.

“We have now put the money in front of these guys. We have put full sovereignty into the stack, the maximum sovereignty you can get in Europe,” said Höttges, referring to the AI industrial cloud in Munich powered by 10,000 GPUs announced by Deutsche Telekom in February.

“Is the government really committed to put … data into these facilities? Defense data? Are they really willing to do so? Are they committing as anchor tenants? And to be honest, I don't see that yet,” stated Höttges. “If we are not seeing the state as anchor tenant of these capabilities, it is not going to work. If they say I can save a cent by going to Amazon or to somewhere else, I can tell you, how can we ever roll scale and get out of this fragmentation which we experience today?”

Last year the EU annonced a European fund of €20 billion to create up to 5 AI gigafactories in the region. Its aim is to “foster public-private partnerships to guarantee a secure investment landscape and nurture a competitive and innovative AI ecosystem within Europe.”

Höttges’ point, of course, is one that DT, Orange, Telefonica and Vodafone have been making for several years: If regulators want European investment in technology, then they need to let European telcos scale through mergers. “Let us gain scale as part of a social contract, and we will invest in technology,” he said.

As the CEO of Eutelsat, Jean-François Fallacher, pointed out during the same panel: “The current geopolitical situation really gives us a lot of weight.” European governments' concerns include falling behind other regions in AI innovation, as well as ensuring data security.

The sovereign enterprise

Geopolitics and regulation are also shaping how enteprises view sovereign AI services.

According to IDC’s data“63% of organizations are now more likely to adopt sovereign cloud services specifically as a result of recent geopolitical events. This sees multinationals sourcing hardware and cloud partners within what IDC calls “sovereign zones”. At the same time, IDC reports the emergence of what it calls “East versus West stacks” at the platform layer. This involves organizations having to “operate across rival spheres without locking themselves into one” so they can ensure data stays within national borders.

Usman Javaid, Chief Product and Marketing Officer, Orange Business, sees a clear trend among its company's international enterprise cutomers for operating multiple geographically defined stacks.

“In the age of cloud we built one single stack for the enterprise to run. Nowadays, in the age of AI, we are building multiple stacks, at least three, one for China, one for Europe, and one for the US,” says Javaid. And the choices that they're making in those stacks are very different”.

Javaid refers to analyst figures that suggest “60% of enterprises are going to build their AI on public and 40% are going to build outside public cloud, and this is our addressable market, the 40%. With geopolitics that 60% may even go down ... and it goes from infrastructure to platform to the solution.”

The challenge for enterprises is that there are "three different types of stacks but not three different types of clients, so how do you ensure a consistent customer experience across them? I believe our role is to help enterprises navigate multiple technology stacks and manage complexity through a trusted platform, ensuring a consistent customer experience.”

SKT is also betting on enteprise growth with its launch at MWC of ‘Sovereign AI Package”. The offer combines AI data center infrastructure, SKT’s proprietary sovereign AI foundation model, which has been adapted to South Korean language and culture, and AI services. Aimed at industry and enterprises, the service is built “on infrastructure that is controlled and operated domestically, taking data sovereignty into account, and providing integrated AI services that have been validated in real industrial settings.”

SKT claims that “telecom operators’ proprietary infrastructure and operational expertise are key to building AI infrastructure and expanding AI services, and that telcos must move beyond the rapid and secure delivery of data to play a leading role in shaping AI infrastructure.”

A need to comply with national sovereign data regulation could play to telcos’ strengths given their history in providing highly regulated national services. Nonetheless, they still have to ensure the demand and pricing warrants investment. Their ability to do so obviously will depend on mix of national competition and regulation, as well as the scale and needs of their existing customer base.

Telenor, for example, was among the companies at MWC making sovereign AI announcements, with the selection of Red Hat to provide a common cloud layer for cloud-native for its sovereign AI factory.

But in an interview with TM Forum towards the end of 2025, Kaaren Hilsen, CEO of Telenor’s AI Factory, was realistic about not only the opportunities but also the challenges of building a sovereign AI business. Telenor deliberately started small with its Nvdia GPU investment and is focusing on both building an ecosystem and establishing the level and nature of customer demand.

“We have got this year and next year to really start bringing in the revenues and prove that the market and demand for sovereign, secure and sustainable AI infrastructure is there,” said Hilsen.