Artificial intelligence (AI) emerged as the biggest acronym from the recent Mobile World Congress (MWC) Barcelona 2025 event, echoing a dominance it held at last year’s annual telecom and acronym fest, but questions continue over how the telecom space will attempt to tame AI's dominance.

Dan Hays, a partner at consulting firm PwC, explained to SDxCentral that while this year’s MWC event did inch closer to its pre-pandemic size, its AI focus surged.

Hays explained that while AI was “blindingly front and center throughout MWC,” the source of that light was questionable. “It's questionable whether everything that had the AI label was truly AI or was really operational at this point, but there was a lot of it there,” Hays said.

This included AI in devices, in the network via radio access network (RAN) equipment, and in customer support systems. However, Hays added that the biggest AI challenge remains a business model that can trigger a new investment cycle from telecom operators.

“I think overwhelmingly, so far, the answer is no,” Hays said.

Hays did note that the event hosted numerous “point solutions” that used AI to solve specific problems. This included pieces of operating support system (OSS) software that can help proactively heal internal network elements.

However, “it doesn't warrant an entire network upgrade just for that capability,” Hays added.

The equipment ecosystem did attempt to use the event to drum up support for the use of AI in RAN equipment, with the most prominent being the Artificial Intelligence Radio Access Network (AI-RAN) Alliance. That group, which was formed just ahead of last year’s MWC event, continues to add members backing its push of using AI in RAN deployments for better performance, lower operating costs, greater efficiencies, and to support new business models.

Analysts have indeed pointed to the benefits AI can bring to the deployment and management of cloud-based open RAN networks, which are more complex orchestration challenges due to the disaggregated multivendor ecosystem. The use of AI could help close performance gaps for open RAN architectures compared with legacy RAN models.

Those benefits are expected to drive an eventual investment push. Téral Research earlier this year predicted the global Artificial Intelligence Radio Access Network (AI-RAN) market, which includes all aspects of AI-driven RAN hardware, software, and services, will grow from $1.7 billion in 2025, to $10.4 billion in 2030, adding that the “current ecosystem is vibrant.”

But not all are convinced.

Chivas Nambiar, GM of telco at Amazon Web Services (AWS), during a press event ahead of the MWC show echoed Hays’ investment concerns, pointing to the still early nature of efforts from ecosystem cheerleaders like the AI-RAN Alliance.

“There’s a fair bit of work in that alliance to try and figure out what the next phase of running RAN functions on top of GPU accelerated infrastructure looks like. I think that’s still pretty early in the conversation,” Nambiar said. “So far, a lot of people are playing around with it, but my expectation, personally, is that as we get closer to that 6G evolution that’s where it’s really going to start to become a decision point of whether we see the cost and performance benefits of the GPU infrastructure running these functions, and that supersedes what we see in custom built accelerator costs.”

Has AI soured open RAN? The nascent nature of telecom AI could also stunt further adoption of broader open RAN systems. While open RAN deployments continue to increase, overall market growth has yet to hit its expected inflection point.

“No one debates the merits of open RAN architecture or of the business model for open RAN,” Hays said. “Right now, the only question for open RAN is one of timing.”

Hays explained that open RAN currently does not have a “strong enough case to warrant large scale rip-and-replace of existing network infrastructure, so as a result the open RAN community is stuck waiting for a natural generational upgrade to take place.”

This sentiment echoed that of former T-Mobile US technology executive Neville Ray, who told SDxCentral at the MWC Barcelona 2023 event the carrier did not push an early open RAN strategy due to timing.

“For us, we didn’t go down on open RAN path when we made our big vendor decisions three years ago,” Ray said at that time. “The pace and scale that we were moving on, features and capabilities we needed, the ecosystem just was not ready. And the O-RAN guys will tell you that.”

Ian Hood, CTO for telecommunications at Red Hat, noted in a press briefing ahead of this year’s MWC event that there also remains some trepidation toward fully embracing any new technology.

“There’s always hesitancy in the market to take on new things,” Hood said, referencing the continued challenges around open RAN technology. “We have to make sure that it’s easy enough to automate and scale and deliver those business benefits to that portion of the network. … We’re going to make this so that it’s easy to add this and add that value to this business. … There is some hesitancy, but we’re working through this, and it’s an ongoing effort for our industry.”

AI is increasingly looking like it might be part of that next upgrade architecture, which could place greater importance to its limelight at this year’s MWC event but might also sour open RAN development and adoption.

“Right now, open RAN is kind of the lemonade stand that’s been set up waiting for someone to drive by and be thirsty,” Hays said.

UPDATE: This story has been updated to provide further clarification on Téral Research's forecasts.