Ericsson and Mobily outline dual AI trajectories for future-proof networks at LEAP

Ericsson and Mobily have used LEAP 2026 to unveil not one broad AI partnership but two separate memoranda of understanding: one centred on putting AI deeper into Mobily’s radio access network, and another aimed at lifting parts of its operating model towards Level 4 autonomy. That distinction, highlighted by Developing Telecoms and TelecomTV as well as Ericsson’s own statements, matters because it shows the pair are pursuing two parallel tracks: intelligence inside the RAN itself,...

and a higher layer of automation across day-to-day network operations. The agreements were announced as LEAP ran in Riyadh from 31 August to 3 September. (developingtelecoms.com)

The first MoU focuses on AI applications in the RAN, where Ericsson said Mobily will test software across RAN Compute, radio units and Active Antenna Systems, while also evaluating centralised use cases delivered through rApps on Ericsson’s Intelligent Automation Platform. Mobile World Live reported that the operator will be given access to Ericsson’s EIAP ecosystem, including more than 100 available rApps from over 100 members, plus software development kits and an AI-powered rApp Developer Assistant. Ericsson said the idea is to let distributed and centralised AI work together, from the network edge to central operations, to improve performance, customer experience, automation and energy efficiency. (ericsson.com)

That RAN agreement is also being pitched as groundwork for services that need faster, more adaptive networks than today’s mobile systems typically provide. Ericsson’s release pointed to future AI-native applications and digital experiences such as AR and VR, autonomous vehicles and so-called physical AI. TLCWorld noted that Mobily’s entry into the EIAP ecosystem gives it more than just a vendor toolset: it opens the door to third-party optimisation apps tailored to Saudi network conditions, a model closer to an applications marketplace than a closed telecoms stack. (ericsson.com)

The second MoU is broader operationally and, for Mobily, probably more immediate commercially. Ericsson said the companies will use Ericsson Operations Engine to prepare and implement use cases in a multi-vendor network, with work spanning operational automation, self-healing, optimisation and energy management. The stated aim is to reduce manual intervention, extract more value from existing infrastructure and explore ways of monetising the network more effectively. Ericsson framed that work around Mobily’s ambition to reach Level 4 autonomy in selected scenarios, building on a partnership the company says stretches back 20 years. (ericsson.com)

What makes the announcements more than exhibition-floor rhetoric is that Ericsson and Mobily have already been testing pieces of this approach in live, high-pressure conditions. Mobile World Live and TelecomTV both pointed back to a 2025 collaboration on explainable AI and an AI application designed to detect and manage 5G uplink interference. Ericsson’s later case study on Hajj 2025 went much further, saying the pair achieved Level 4 automation for uplink optimisation during the pilgrimage season through closed-loop controls. According to that case study, the system handled more than 23,000 AI-driven decisions per hour, executed over 15,000 automated corrective actions before the event, introduced more than 50 new KPIs for live failure monitoring and carried out 598 automated corrective actions across 116 sites. Ericsson said the result was a 15% improvement in SINR, a 9% gain in higher-order modulation efficiency, a 7% rise in spectral efficiency and a 10% uplift in 5G throughput. (mobileworldlive.com)

The companies also carried that work into Hajj 2026. In a May 2026 statement, Ericsson said Mobily more than doubled 5G capacity in the Mashaer area, expanded coverage using additional sites and new spectrum, and scaled zero-touch AI tools for uplink interference optimisation and coverage-gap mitigation. The same update said the partners added an AI-based end-user monitoring tool and a GenAI assistant for operations teams. Mobily’s chief technology officer, Alaa Malki, said the companies had “evolved how we plan, optimize, and operate the network” during the pilgrimage season, while Ericsson’s Håkan Cervell described Hajj as “one of the most demanding connectivity environments in the world.” (ericsson.com)

The near-term business case may lie less in futuristic applications than in power bills. TLCWorld argued that energy management is the likeliest early win because the RAN is one of the most power-hungry parts of a mobile network, and AI-led automation can adjust capacity and radio resources cell by cell as traffic rises and falls. That reading fits Ericsson’s own emphasis: energy efficiency appears in both the AI-RAN and autonomy announcements, alongside the longer-horizon language about new digital experiences and network monetisation. (tlcworld.it)

There are still important caveats. These are memoranda of understanding, not disclosed supply contracts, and neither Ericsson’s releases nor the trade reports that followed set out spending commitments, deployment deadlines or specific performance targets for the new programmes beyond Mobily’s Level 4 ambition in defined use cases. Even so, the package of announcements suggests a clear direction: Mobily wants AI to sit both inside the radio layer and above it in operational control, while Ericsson is using Saudi Arabia as a showcase for how rApps, closed-loop automation and multi-vendor orchestration might move from special-event pilots into routine network management. (ericsson.com)

Source: Noah Wire Services