MWC will probably be an AI show

The MWC trade show in Barcelona, Spain, is just a few weeks away, but it will probably go down as the show that was dominated by AI.

Already companies across the wireless industry are leaning heavily into the intersection of AI and telecom. In some cases, that means putting AI services onto smartphones and other gadgets, in others it means applying AI technology into network operations. In either case, network operators are hoping for profits. Or at least smaller losses.

"Artificial Intelligence provides the telecoms industry, and the societies it serves, with huge opportunities to launch new services, improve connectivity and customer experience. Overall, it's estimated that AI could contribute $15.7 trillion to the global economy by 2030," said Alex Sinclair, the GSMA's CTO, in a release Tuesday. GSMA is the association behind the annual MWC trade show. "However, it's critical that AI is democratized to ensure that all parts of the connectivity industry and their customers, wherever they are in the world, benefit."

Along those lines, GSMA this week announced a deal with IBM to launch an AI training course through the association's GSMA Advance program. The training program will "span a wide range of topics, from fundamental AI principles to specialized Gen AI applications in telecoms," and it will offer access to the watsonx AI platform that IBM began discussing at last year's MWC show.

To be clear, GSMA isn't the only telecom player to tout its AI chops. Also this week, Cradlepoint, Juniper Networks and others made AI-related announcements. 

AI in the network

For Cradlepoint, the company on Tuesday announced "new comprehensive AI functionality to make 5G networks smarter, simpler, and more secure." Specifically, the company said it would use AI to prevent data loss, to enhance network performance, and to provide network traffic analysis.

"Our AI strategy exemplifies our commitment to accelerating the deployment of 5G for businesses by enhancing the visibility, manageability and performance of the cellular network," said Donna Johnson, Cradlepoint's CMO, in a release.

The move is noteworthy considering 5G equipment giant Ericsson closed its purchase of Cradlepoint in 2020. The move bolstered Ericsson's pursuit of the private wireless networking sector.

Separately, Juniper said it expanded its own use of AI technologies to improve operations in the network and the data center. Juniper also made sure to note that its AI technologies have been "trained on seven years of insights and data science development." That's a nod to the AI technology that helped drive HPE's proposed $13 billion takeover of Juniper, announced earlier this month.

And the announcements by Cradlepoint and Juniper are likely just the start. For example, Deutsche Telekom's Alex Jinsung Choi, chair of O-RAN Alliance, has been posting extensively on social media about the role of AI in enterprise operations and radio network management.

Finally, AT&T this week made sure to tout its own AI operations in an announcement that it plans to expand its AT&T Ventures funding organization. "AT&T helped launch the field of 'artificial intelligence' in the 1950s, and today we have deployed our generative AI tool called Ask AT&T to more than 60,000 employees," the company wrote.

It's unclear whether new AI guidelines from the Biden administration speed or slow such efforts.

Preparing for Barcelona

The newest AI announcements help to set the stage for the upcoming MWC show. AI will probably be the buzzword to dominate MWC Barcelona in the same way that virtualization, cloud, open RAN and private wireless dominated past shows.

Already AI is at the center of a wide range of events scheduled at MWC, including a Wednesday panel titled "Is 2024 the year of the AI-centric telco?" and a Monday event titled "Harnessing Gen AI at telco scale."

Another session, "Developing Gen AI that cares," may also address ongoing questions about the effect AI will have on hundreds of thousands of workers in the telecom sector. Already operators like AT&T have suggested that AI is helping to reduce call volumes in customer care call centers. It's also making the operator's field operations, consulting and contractor expenses more efficient – efforts that dovetail with AT&T's ongoing efforts to reduce its spending and boost revenues.

But what will AI mean for core telecom operations, including network traffic?

In the data center and on the phone

First, it's clear that actual spending on AI services remains focused squarely on the data centers that house cloud AI services from the likes of Amazon, Google and Microsoft. Specifically, the financial analysts at TD Cowen wrote that demand for high-performance data centers – ones that can handle the development of AI technologies – remains sky-high in the early parts of 2024.

"We believe the stage is set for record 2024 hyperscale data center leasing driven by AI," the TD Cowen analysts wrote in a recent note to investors. "Our checks point to a year-over-year doubling of Google's data center capital expenses. Meta is back in full swing post-pause while Oracle ramps deal sizes notably. However, Microsoft continues to lead the leasing charge with Amazon looking to close the distance."

And that rise in data center spending, they noted, could be substantial enough to put a strain on the nation's power grid. "As data center demand ramps, we expect the challenges associated with procuring utility power to become more pervasive," they wrote.

Indeed, they cited warnings from US utilities like Dominion Energy and Georgia Power that they may not be able to support the proposed construction of more power-hungry data centers. "To satisfy the demand for power in Atlanta, the utility would need to double its capacity in just three years," they cautioned.

But Samsung's newest Galaxy smartphone offers a bit of hope to telecom network operators jealous of all that spending on AI computing power. The gadget boasts a wide range of AI-powered services, including transcribed summaries of voice recordings, complete with subheads and timestamps.

As The Verge notes, those summaries are derived from a cloud computing connection rather than calculated on the device itself. Meaning, users are going to need a speedy wireless Internet connection to make all of their Galaxy AI services work – and that's undoubtedly music to the ears of mobile network operators around the world.