Ericsson has introduced an AI-first approach to building mobile networks, unveiling a new portfolio of radios, antennas and RAN software designed to support the rapid growth of AI-powered devices and applications. The launch comes ahead of Mobile World Congress 2026 in Barcelona, where the company is expected to showcase how artificial intelligence is reshaping network performance, efficiency and monetization opportunities for telecom operators.
Research firm Dell’Oro Group said the size of the global RAN market in terms of revenue will remain mostly flat in 2026. The top 5 RAN suppliers are Huawei, Ericsson, Nokia, ZTE, and Samsung. In 2025, Huawei and Nokia gained ground, Ericsson and Samsung were stable, and ZTE’s RAN revenue share fell.
AI-driven networks designed for new traffic demands
The surge in multimodal AI and augmented reality applications is driving new network requirements, particularly for uplink capacity, low latency and real-time optimization. Ericsson’s latest portfolio aims to help communications service providers monetize AI devices with differentiated connectivity while reducing total cost of ownership and improving energy efficiency.
The company’s new AI-ready radios incorporate Ericsson Silicon with neural network accelerators, enabling on-site AI inference in Massive MIMO radios. These programmable matrix cores are integrated into the Ericsson Many-Core Architecture and optimized for machine learning workloads, allowing real-time network optimization across the full RAN stack.
Marten Lerner, Head of Networks Strategy and Product Management at Ericsson, said the new portfolio represents the company’s “final step to full AI enablement” across its network offerings.
Ten AI-ready radios target uplink and efficiency gains
The hardware lineup includes ten AI-ready radios designed to improve downlink efficiency and deliver major uplink performance gains for AI and AR applications.
Key highlights include:
High-power FDD Massive MIMO radios such as AIR 3286 and AIR 3211 supporting both TDD and FDD deployments.
Expanded eight-receiver portfolio with Radio 4891 and Radio 4458 to boost uplink capacity.
High-power triple-band radios Radio 4488 and Radio 4464 for network consolidation and RAN sharing.
TDD Massive MIMO advances including AIR 3267 with 600 MHz bandwidth in a 13 kg form factor and AIR 6492 featuring 480 W power and 256 antenna elements.
These innovations aim to address rising data traffic while enabling operators to deploy AI capabilities directly at the network edge.
AI RAN software introduces real-time intelligence
Ericsson is also rolling out new RAN software features that use AI to improve reliability, efficiency and latency performance.
The software suite includes:
AI-managed beamforming to optimize radio performance dynamically.
AI-powered outdoor positioning.
Instant coverage prediction powered by advanced AI models.
Latency Prioritized Scheduler and Low Latency Mobility delivering up to seven times faster response times.
These capabilities complement Ericsson’s AI-native Link Adaptation software and enable new service offerings previously not possible in mobile networks.
New antenna portfolio improves spectrum efficiency
The portfolio also features five high-performance antennas designed to simplify site deployment and enhance spectrum utilization. Energy-efficient passive antennas built on trio net design aim to maximize uplink performance, carrier aggregation and spectral efficiency. The expanded interleaved AIR portfolio now includes three new configurations to support flexible TDD and FDD Massive MIMO deployments.
VodafoneThree highlights real-world deployment impact
VodafoneThree is already working with Ericsson to integrate AI across its combined 5G infrastructure.
Iain Milligan, Network Development and Infrastructure Director at VodafoneThree, said the integration of AI across the network will reduce operational complexity, boost energy efficiency and deliver faster, more reliable connectivity.
Enabling AI monetization and network efficiency
Ericsson says its AI-powered hardware and software portfolio enables operators to deliver differentiated user experiences, capitalize on demand for AI-enabled devices and accelerate time to value. By embedding AI deeply into the RAN, the company aims to help telecom operators support emerging AI workloads while lowering operating costs and improving network sustainability.
BABURAJAN KIZHAKEDATH