Carriers Embrace AI and Move to Cloud, but Questions Remain
AI will mean “fundamental changes” to the way carriers build and operate their networks, Deutsche Telekom’s Franz Seiser said Tuesday during a TelecomTV digital signal processing forum, streamed live from Windsor, U.K. Seiser and other speakers agreed that most carriers are making their first, still-tentative steps to embrace AI.
Sign up for a free preview to unlock the rest of this article
Communications Daily is required reading for senior executives at top telecom corporations, law firms, lobbying organizations, associations and government agencies (including the FCC). Join them today!
Carriers must begin their AI journey by ensuring their operational data is available in the proper format and architecture, said Seiser, head of Deutsche Telekom's technology, development and architecture team. “If you’re talking about 6G or the current G or whatever G, it’s all about the data architecture you need to have in place first,” he said. “Then the opportunities are limitless.”
Seiser said his company is making its entire core network cloud native, building its cloud level as the bottom level and automation as the upper level. The carrier takes its central monitoring system from vendors, “but we do not take the management software,” he said. “We ask them to integrate into our pipes” and automation tools “so that we have a consistent automation layer on top of all our network functions.” By the end of 2028, nearly all the network will be in the cloud, he added.
Moving to the cloud makes network management cheaper, faster and more agile, Seiser argued. A change that would take weeks or even months on a legacy network can be completed in a matter of days using cloud-based technology, he said. “There’s a tremendous difference in speed.”
Chris Meads, commercial director at Vodafone Intelligent Solutions, said businesses are “struggling” to find “value at scale” from generative AI -- research shows only 13% are getting “significant value” from it. Companies are spending three times more on technology than the process and staff needed to implement AI, he said. “We need to get away from this culture of constant proofs of concept” and get to “driving significant value at scale."
Manish Singh, chief technology officer of Dell Technologies’ telecom systems business, said AI-native means that a system was built “with AI embedded and as a foundational technology, not something that’s layered on.” Since 5G networks are mostly rolled out, “what we’re going to see is a lot more of the AI pieces coming on top of it.”
Singh said most telcos are getting started in AI to improve customer service and their networks. As work on 6G starts, he said, there’s a consensus that the next generation of networks will be AI-native. “You can go in many directions with AI,” Singh said. “It’s a foundational technology.” Questions include “what’s the business value? What’s your unique differentiator? Where should you focus?”
Carriers worldwide are experimenting with AI and “scaling up smaller efforts,” said Gabriela Styf Sjoman, BT Group managing director-research and networks strategy. “If we want to go beyond that … we need to address [AI] holistically, and that requires that we reassess the operating models.” There’s “great potential” for AI, but work remains.
Hanen Garcia, Red Hat chief architect-telecommunications, said we have yet to see how AI will transform the telecom industry. “We are at the very beginning,” he said. AI “will be everywhere” eventually.
The “cloudification” of networks means “faster time to go to market, more agility [and] operational efficiency,” said Abir Hossain, vice president-technology and operations at digital signal processing company Elisa. Companies need to make sure the move to the cloud is “not just a lift-and-shift” project and is “actually transformational,” with real-world business impact, he said. “It’s not just an IT project; it’s a business-transformation project.”