- Ericsson sees AI as both a tool to improve today’s networks and a driver of future architectures
- The company is pushing AI deeper into network operations, while urging caution around trust and autonomy
- Ericsson views 6G as AI‑native, built to support machines, sensing and physical AI
MOBILE WORLD CONGRESS 2026, BARCELONA — At Mobile World Congress 2026 in Barcelona, Ericsson leaned into that phrase we keep hearing everyone using for their AI strategy — AI for networks and networks for AI. The vendor says that phrase will define the next phase of telecom evolution: using AI to improve networks today while prepping networks to support the demands of AI-driven applications tomorrow.
In an interview at the MWC 2026, Ibrahim Eldeftar, global head of solution line cognitive software and services for Ericsson, told Fierce Network that the vendor's approach is, in-fact, “AI for networks and networks for AI," noting that the two priorities are inseparable as operators confront rising complexity across radio access, core networks and operations.
“We see AI as top priority in our strategy,” Eldeftar said. “That means preparing networks and connectivity for the workloads of AI, but also using AI inside our own portfolio to make networks better.”
'AI for networks'
On the operational side, Ericsson is pushing deeper use of AI across radio, core and network management layers. The company highlighted long-running investments in AI-driven radio features, including link adaptation and spectrum optimization, which it says are delivering double-digit throughput gains in live networks.
Ericsson is also pushing AI into what it calls the intelligent core, using machine learning to improve resiliency, service quality and operational efficiency. A key focus is autonomous networks, where AI can detect issues, identify root causes and resolve problems with minimal human intervention.
“The question is how you achieve autonomy, from detecting issues in the network to finding the root cause and resolving it without human intervention,” Eldeftar said. “These are not theoretical examples. These are live deployments with customers.”
Still, Eldeftar said that trust remains a barrier, particularly when AI is applied to national infrastructure. While the technology is largely ready, operators must be confident that automated systems can make changes safely and predictably.
'Networks for AI'
Looking ahead, Ericsson sees AI as a defining force shaping both 5G evolution and early 6G design. The company is working with operators to make 5G and 5G-Advanced more spectrally efficient, freeing up capacity while laying the groundwork for future generations.
Ericsson also views 6G as “AI-native,” with intelligence embedded directly into network functions and exposed through APIs. Eldeftar said future networks will need to support not only human communications, but also machines, sensors and what Ericsson refers to as physical AI.
“We see 6G as AI‑native,” he said. “Functions that used to sit in external systems will be embedded directly into the network, with AI at the center of the architecture and exposed through network APIs so customers can use AI capabilities directly from the network itself.”
“Some of the use cases in 6G are not exactly human,” Eldeftar noted. “We start to count machines. We talk about integrated sensing, physical AI and new types of devices.”
Rather than replacing existing generations, Ericsson expects 5G and 6G to coexist for years, with AI acting as the connective tissue that allows networks to evolve continuously.
“Connectivity demand will continue,” Eldeftar said. “And AI will be central to how we meet it.”
Read all of our coverage from Mobile World Congress 2026 in Barcelona here.