Nothing you saw at MWC 2025 is going to change the trajectory of the MNOs’ consumer businesses. There are no new substantial consumer revenues to secure, bar a renewed push in re-connecting cars.
They have already lost the enterprise 5G market. The single throat they need to choke is not a major vendor; it is their own – specifically the weight of bureaucracy that slows these operators to a crawl. Without cutting out the rot in their operations, these MNOs are going to shortly turn septic.
Rising opex and capex, stagnant prices, poor return on capital invested, and geopolitical turmoil have swiftly shattered market confidence. Consumers have all they need from the MNOs, and will not pay more for their services. The current industry hope is Network API, a trend that requires significant investment with very unclear returns – a losing strategy.
But AI really will disrupt things. You must tolerate the marketing hype, but the new crop of agile vendors that hope to outpace the incumbents will use LLM-based automation to tackle the legacy technological debt – unwinding their labyrinthine IT and BSS networks.
This is the crux of the matter. With no outward hope, the MNOs will have to look inward. These new tools will offer millions of potential job cuts across the industry – a win for the cost base, but a potential PR nightmare. In both the IT and OT teams, a huge number of roles are on the near-term chopping block.
The loss of institutional knowledge, as skilled staff retire, churn away, and are eventually replaced, will leave MNOs extremely dependent on black-box implementations from vendors. Gone are the days of a senior engineer understanding the entirety of network operations. The new age will be automated, but never fully understood – not that we can reliably claim that the leadership in these MNOs understand their networks today.
Lost at (the Balearic) Sea
It was a Mobile World Congress without a central theme; and an industry unclear on where it should be looking for growth. AI was festooned on many booths, but was all too often a solution looking for a problem. Network APIs were present but mercifully muted, and 6G was largely absent. Sustainability was disappointingly missing, and Open RAN was unsurprisingly quiet.
It is clear that AI threatens to disrupt the incumbent vendors, but like every step in the mobile stack, there is a very sticky vendor looking to keep those doors slammed shut. However, these operators are incapable of new revenue growth in the mobile market, for a multitude of reasons (a legacy of costly mistakes), and so to improve their financial position, they must look inward – and prepare to wield the knife.
This was the largest MWC to date, with some 109,000 attendees across the four days. As usual, the last day sees a lot of the executives slink back home, and the show floor opened up to local students, which helps to pad the numbers. It felt like the busiest MWC so far, but was well run – besides our usual press facilities gripes.
Most of us can remember the successive topics that the marketers have jumped on. Following the 2008 financial crisis, the first major new growth area identified by this industry was the IoT. We know how that panned out, in terms of meeting the promises of billions of lucrative connections for the MNOs.
Next was blockchain, which morphed into NFTs, and then the first wave of 5G hype kicked in – focused on those future-looking use cases, most infamously robotic surgery. When this began to falter, we shifted to AI, and now the Network API angle is using up most air in the room. Somewhere in there, sustainability began being talked about, but as seen at MWC 2025, this has fallen quiet.
In parallel to these new opportunities, the MNOs carried on underserving their core consumer customers. For each dropped call, billing mistake, or coverage gap, a decision maker inside an enterprise risked taking a dim view of MNO capabilities. These problems have damaged the MNOs’ abilities to break into the private networks market, and continue to dent uptake of additional services, including Mobile Edge Compute (MEC).
Having overpaid for 3G spectrum, the MNOs were somewhat forced (outside of the usual amount of greed) to overcharge for 3G services like MMS and video calls. This opened the door for OTT rivals in the 4G era, which saw control of communication services wrenched from the MNOs. This forced their hand into prioritizing data in their tariffs, and in this race to the bottom, the chance of growth collapsed.
Unlimited data bundles carried over into 5G, and now there is nowhere to go. Network slicing will leave a bad taste in the mouth, and with so little to differentiate the MNOs from their main rivals, it is no surprise that more consumers are moving to cheaper MVNOs. The price difference between these two options illustrates the scale of the bloat inside the MNOs.
If 80% of IT budgets are being spent supporting legacy systems, it is no surprise that the MNOs are struggling to grow with that 20%. Internally, they face a staffing crisis, as their most experienced engineers age out of the MNO workforce – and are replaced with staff that do not fully grasp the operation of these messy networks, and who are accustomed to relying fully on vendor tools.
But here is the glimmer of hope. Automation, leveraging advances in LLM-based integration, could dramatically reduce the number of staff required by the MNOs, in both network engineering and customer support.
However, laying off millions of staff is a PR headache, but given the way the world is turning, with regard to the national importance of these networks in this geopolitical turmoil, these ‘too-big-to-fail’ MNOs have a get-out-of-jail-free card to play – cozying up to governments to acquire a slice, should things get too tough.
Out network divestment pitch continued to turn heads at MWC. Separating the customers from the network, via an MVNO spin-out, would vastly simplify matters. If MNOs are already comfortable selling off their tower assets, the next logical step in that journey is selling off the RAN. If you’ve sold the RAN, why would you hold onto the core, and so on and so forth.
A new concern appeared at MWC 2025, which was that for many of the future-looking edgy AI compute workloads, a chronic lack of power at many of these telco sites is rearing its heady. This is an opportunity for new appliances that can reduce the power usage of core MNO functionality, but a real problem for the GPU-aaS pitch from Nvidia, as these monstrous GPUs might not have enough available electricity on hand. Upgrading this availability takes a very long time, we are told.
Our most outlandish idea caught the few we brought it up to off guard. If we were ruthless private equity buyers, looking to acquire distressed assets on the cheap, we would work to devalue the balance sheet valuations of these spectrum licenses, which are propping up MNOs the world over. Given the new political reality in the US, that is certainly market ripe for the likes of a16z.