There was an undoubtably a more upbeat tone at Mobile World Congress this year, punctuated by a post-Covid record crowd of 109,000 people attending this goliath communications trade conference.
From this analyst’s point of view, there were two macro themes fueling the optimism: 1) A comeback in capital spending as 5G standalone (SA) deployments reach critical mass; and 2) The excitement about AI and what it means for telecommunications.
As usual, some of these excitement levels may be unfounded, but it’s undoubtably better than the slower macro environment suffered by the telecom market over the last couple years. Let’s dive into some of the details. As the table below shows, capital spending is set to bounce back this year.
Telecom capex bounced back in 2024 from 2023 levels and is expected to rise modestly as S&P Global expects it to grow 3% in 2025.
Nokia Highlights Growth
At a pre-MWC kickoff conference I attended hosted by Nokia, outgoing CEO Pekka Lundmark dove right into the optimism, pointing to Nokia’s comeback with the renewed surge in spending.
“The markets are coming back,” said Lundmark. “We had 9% growth in Q4.”
Lundmark, who will soon retire to make the way for new CEO Justin Hotard, pointed to Nokia’s deployment of 120 standalone core networks across the world. He also believes “it is still a fairly early stage.”
Lundmark thinks the industry can not only further benefit from the deployment of 5G and its synergies with AI, but also from cost savings due to AI-based automation. He sees “agentic AI across the portfolio,” saying the company can recognize hundreds of millions in savings in automation for customers.
Michael Clegg, VP and GM of 5G and Edge for Supermicro, told me a similar story. He believes a lot of the 5G story was derailed by early hype and deployment of hybrid 5G architectures that were not really 5G.
"5G with nonstandalone confused the market,” said Clegg. “It was really 4G+. Real 5G is 5G standalone. So, in a way, 5G is only coming to market now. The benefits are starting to show up, it's started with 5G SA."
T-Mobile’s success story
Another lesson from the telecom rebound may be about focus and execution. Although many in the industry have grumbled about slow growth as they jealously eye the boom in AI and cloud, it’s a fact that some providers have thrived.
T-mobile is one of them, as witnessed by the stock chart below. Shares have more than doubled in the past four years.
At the Nokia event, which had a panel of customers, T-Mobile President of Technology Ulf Ewaldsson talked up T-Mobile’s success in North America, where it has undeniably had the most financial success of all the operators.
“We had 5G SA Standalone, and we used 5G to lead the competition in the United States,” said Ewaldsson.
T-Mobile’s focus on spectrum, fixed-wireless access, and customer service paved the way to this success. With all the chatter about network slicing, autonomous driving, and IOT, fixed broadband access has in fact been the dominant use case for the nearly 200 services tracked by Futuriom’s 5G Services index, with them featuring as the primary use case in 120 of the 200 services, as seen below.
AI at the Edge
For AI in telecom, the optimism shifts to the AI-at-the-edge story, led by the twin benefits of optimization with AI-RAN and other potential new applications for AI.
At a pre-event conference hosted by Nokia, Ryuji Wakikawa, head of Softbank’s Institute of Advanced Technology, said AI inference offers a golden opportunity for telecom providers.
“I think datacenter for training is an investment,” said Wakikawa. “More important for operators is inference. Where does the revenue come? We want to capture that revenue. We need technology and that’s AI RAN.”
A recent Rosenblatt report on top AI startups highlights the same trend: What’s next is inference, and what does it mean for enterprises?
“The anticipated growth of edge AI represents a significant opportunity” write the Rosenblatt analysts. “There is an opportunity for technology disruption and a shift in leadership as the focus shifts from the datacenter toward applications and inference at the edge.
As telecoms build out edge infrastructure with AI gear, inference represents just one potential application. AI-RAN technology was another big theme at MWC.
AI-RAN uses AI to optimize radio-access networks. It’s a telecom-specific application. The partnership known as the AI-RAN initiative, part of the AI-RAN Alliance, is promoting the integration of AI technologies into 5G and 6G radio access networks to optimize performance, automate resource management, and enhance overall efficiency.
AI-RAN and other applications were a big of focus of a flurry announcements and partnerships too numerous to name. But in addition to some of the announcements I’ve mentioned from Nokia and Softbank, other companies looking to log onto the theme included Amazon, Arrcus, Orange, NVIDIA, and Samsung. In fact, I’d go as far as to say the number of AI-RAN announcements was enough to make an analyst’s head explode. It would have been more refreshing to see more news about enterprise AI applications at the edge.
Supermicro’s Clegg told me that the AI opportunity for telecoms at the edge of the network is large, as enterprise demand for inference models can broaden the market.
“A lot of money going into the training side, but AI becomes more interesting when you do inferencing at scale,” said Clegg.
In conclusion, this year’s MWC was huge and a bit messy, but at least the backdrop of growth permeated the scene. For the next couple years, it’s all going to come down to how telecom can monetize the enterprise and consume demand for AI, a topic we’ll be studying closely.
Scott Raynovich has been following Internet and communications technology long enough to remember what Token Ring and Silicon Investor were. A technology analyst and writer for 30 years, he helped champion the rise of the Internet bubble as New York Bureau Chief for the seminal Red Herring magazine during the 1990s, embraced the telecom boom and bust as Editor in Chief of communications trade Light Reading for eight years in the 2000s, and published one of the first reports on SD-WAN technology at the Rayno Report in 2014, which was acquired by SDxCentral.com in 2015. After doubling the research revenue at SDxCentral.com, Raynovich left to start a new research business in 2016. Futuriom was founded in 2017 to deliver insights into the hottest trends and demands driving cloud technology.
Op-eds from industry experts, analysts or our editorial staff are opinion pieces that do not necessarily represent the opinions of Fierce Network.