How AI is redefining mobile networks
AI is no longer a future aspiration, it’s the engine driving network performance, programmability, and innovation. At Ericsson, we’re not just exploring — we’re deploying. Our AI RAN solutions are live, delivering measurable improvements in performance, energy efficiency, and network automation.
But what does this actually look like in practice? Let’s explore the evolution of AI in networks and what we uniquely bring to the table.
AI’s dual role in mobile networks: What Ericsson brings to the table
AI plays a dual role in today’s mobile networks:
AI for Networks: Optimizing throughput, reducing latency, improving energy efficiency, and lowering operational costs.
Networks for AI: Enabling scalable, high-performance infrastructure to support AI-driven applications.
We take this duality a step further by making AI foundational to the network, embedding it into the architecture itself rather than adding it on top.
What sets us apart? Here are our four key pillars:
Intent-Based + AI-driven Networks: We combine intent-based networking with embedded intelligence to create programmable, high-performing networks that are both flexible and scalable.
Support for open architectures: Compatible with both purpose-built and cloud-native RAN deployments, allowing service providers to select their hardware vendor of choice.
Hardware and software co-design: AI is not just software — it’s deeply integrated into the RAN hardware using custom silicon and model compression to support real-time inference and execution with microsecond-level latency.
Flexible AI deployment architecture with rApps and Ericsson 5G Advanced solutions.
🔍 As Melike Erol-Kantarci, our Head of AI RAN, explained in her Mobile Europe interview, “AI is not a built-on. It's foundational.”
Insights from the AI Summit: What the industry is saying
At the recent Light Reading AI Summit, thought leaders across telecom aligned on three core themes:
Balance: Between real-time (embedded) and non-real-time (centralized) AI applications.
Challenges: AI model lifecycle management, energy efficiency, and cross-vendor interoperability.
Vision: Towards autonomous, AI-native RAN.
Speakers emphasized the importance of balancing business cases when deciding between embedded and centralized AI — and highlighted that deep telecom expertise is essential to developing effective AI algorithms.
📌 Don’t miss it: Watch the AI RAN On-Demand Webinar to go deeper into our technical strategy.
🎧 Bonus: Check how Bell Canada achieved a 20% improvement in downlink throughput and 10% spectral efficiency gains using Ericsson’s AI-native link adaptation.
Looking ahead: AI in 6G
In 6G, AI will serve as a key and native component, shaping the network's architecture, capabilities, and services. We’re doing this in close collaboration with the ecosystem and leading industry players.
Ericsson’s portfolio is uniquely built to meet both near-term demands and smooth evolution to 6G. This includes:
Ericsson RAN Compute, engineered for AI performance with a hybrid architecture that integrates third-party CPUs, in-house ASICs, and Ericsson Many-Core Architecture (EMCA) with a massively parallel processing capabilities for fast AI inference and execution — achieving best-in-class processing for AI workloads
Ericsson 5G Advanced software, enabling microsecond-level execution of time-critical AI functions at the radio site
Ericsson rApps and Open SMO, supporting centralized AI use cases such as network-wide optimization, deployment orchestration, and intelligent healing
💡 Watch Episode 3 of our Tech Unveiled series, "Unlocking AI for Networks with 5G Advanced", to dive into Ericsson solutions.
How are you thinking about AI in your network evolution? Let’s talk about how we can help you accelerate that journey — with real deployments, proven results, and a roadmap to 6G.
Account Director | Driving business outcomes through technology
1moA Iran 🇮🇷...can't unsee that
@ Poonam Parihar The Feedback Loop
Brand Alchemist • Chaos Whisperer • Soul Scout • Thought Twister • I hunt the truths the surface hides! Shikamaru Nara with destiny as my guide.
1moAI’s not a patch, it’s the pulse in the wire, Making networks smarter, faster, and flyer. From RAN to the edge, they’re setting the pace, Ericsson’s tech? A whole new race!