MWC 2025 Recap
It is almost an unwritten rule among MWC participants to rave about the tapas in Barcelona when discussing the show in a report/blog or conversation, so let’s follow the tradition and give a shout-out to the amazing tapas and sangria we had in Barcelona. Now that’s out of the way, let’s delve into what MWC 2025 was all about. The premise of the show, as in the previous year, was all about the integration of Artificial Intelligence (AI) in telecom. Here are some major themes and trends we observed at the event:
AI-powered Smartphones: While MWC 2025 has shifted from being the primary venue for flagship device releases, major Chinese brands like Xiaomi continue to seize the opportunity to showcase their latest models. Xiaomi released its 15 Ultra flagship smartphone that stands out with its 200-megapixel telephoto lens. However, the phone’s most captivating feature is the separately sold Photograph’s Kit, which transforms the device into a Point-and-Shoot camera. Like other Android flagship devices, the 15 Ultra supports Gemini Live AI and Xiaomi’s exclusive AI skills.
HMD captured the spotlight with pre-show announcements including a partnership with FC Barcelona, a globally recognized soccer team with a substantial Hispanic community following in the US. HMD also released a featurephone with the hope of attracting youngsters interested in digital detox. However, our data consistently indicates a lack of interest in featurephones. Notably, HMD unveiled the Fusion X1, a phone specifically designed for teenagers, with parental controls that will appeal to parents looking to limit their children’s phone usage. This niche is crucial, as Circana’s previous survey revealed that many parents find the usage limits provided by operating systems insufficient in monitoring their kids’ smartphone usage.
Another brand making an impact outside the show venue was Nothing, which introduced its latest phones, the 3A and 3A Pro, which will compete against the affordable Galaxy A series or Pixel’s affordable “a” models in global markets, though the U.S. market impact will be marginal as the Nothing phones operate only on T-Mobile’s 5G network.
Democratization of AI: None of the aforementioned phones released at MWC 2025 will have a major impact on the U.S. smartphone landscape, which is heavily dominated by Apple and Samsung. Samsung used the venue to introduce its latest mid- and low-tier Galaxy A series models, the A26, A36, and A56 models, which boast improved specs including AI support over their predecessors released last year. The inclusion of AI support in these mid-tier models, like Apple’s affordable iPhone 16e, is a sign of the democratization of AI on smartphones; this was an expected outcome as, apart from Samsung Galaxy S24’s short-lived success as the first AI-powered smartphone, brands have had difficulty using AI features to entice users to upgrade to new flagships.
Coincidentally, Apple just recently announced delays in the rollout of some of the GenAI features that were heavily marketed as part of Apple Intelligence. It is true that Apple is behind the curve with its AI efforts, especially around building the ecosystem around its AI agent, but these delays on the Apple Intelligence front should have limited impact on iPhone upgrades and churn. Apple has an incredibly high retention rate for its iPhones, and 90%+ of iPhone users do not jump ship and get an Android phone instead. While novelty features such as Apple Intelligence trigger consumers to upgrade to a new iPhone, most iPhone users have a firm upgrade cycle, which is mainly dictated by the battery health of their phone.
Agentic AI: One of the most interesting announcements at last year’s MWC came from Deutsche Telekom, T-Mobile USA’s partner company, when they introduced the AI phone concept. The concept is based on an AI assistant/agent taking over all the phone tasks without the need for the user to interact with the applications installed on the phone. The telecom giant expanded upon this concept at MWC 2025 and announced a partnership with the Perplexity AI platform. The AI phone is expected to be introduced later in the second half of the year and commercially debut in 2026 for a price tag of under $1000. Despite laying (and continuously evolving) the foundations of communications, carriers have long been criticized for having their networks turn into “dumb pipes,” allowing tech giants like Google and Apple to take over customer interaction and control. The development of an OS-agnostic Agentic AI phone has the potential to disrupt the mobile device landscape like Apple did with the iPhone and App Store debut back in 2007, but it would be naïve to expect app ecosystem players to give up easily. We have already seen brands such as Honor discussing the Agentic AI interface taking over the app-based interface consumers have been accustomed to since the invention of smartphones, but the new interface will still rely on apps running in the back-end.
Responsible AI: Generative Artificial Intelligence (GenAI) has been a key selling point for AI-powered devices and services since the AI concept’s mobile debut. However, the rapidly evolving GenAI capabilities have also raised concerns about deepfake content misuse for influence. It was refreshing to see vendors such as Honor address this security concern with its AI deepfake detection software, integrated into its Alpha Plan program, which involves a $10 billion investment in exclusive AI services. We’ve also seen companies like Spain’s Telefonica emphasize the importance of responsible and ethical AI services, and we anticipate these references in future AI marketing from telecom companies.
AI here, AI there, AI everywhere: AI permeated every aspect of the show, as expected, but it was also noticeable that the AI reference was overused in some instances. Telecom companies have been developing machine-learning programs to streamline their network and service processes for years. Now, they simply slap “AI” at the front or end of the process name. Despite this, it was refreshing to hear global telecom companies discussing the advanced state of their machine-learning (ahem, AI) processes, which range from network optimization to APIs.