Dell’Oro Group Finds 6G RAN Planning Is Shifting From Unknowns Toward AI and Economics

Dell’Oro Group says 6G RAN planning is moving from transformative narratives toward deployment realities, with evolutionary wide-area models, AI-native RAN designs, and open fronthaul expansion among areas showing clearer alignment.

Market Overview

The note frames 6G RAN as less reliant on what remains uncertain than earlier 5G cycles, with a focus on deployment efficiency and economics rather than new revenue narratives. It also characterizes baseline 6G as evolutionary, with specific RAN elements more likely than others.

It states that wider-spectrum channels over existing macro grids remain the main mechanism for increasing capacity and lowering cost per bit in 6G. The document also points to small cells as important for densification and indoor coverage, with early focus expected to prioritize outdoor macro layers.

Key Findings

Dell’Oro Group estimates AI and 6G should improve efficiency, with gains in the “10% to 50% range.” It ties the economics to a RAN cost share of “less than 15% of the overall wireless capex and recurring site opex over the life of the cell site,” arguing that new-site requirements can break cost assumptions.

The note describes an investment-logic shift from 4G and 5G’s technical success and capacity gains to a 6G focus on efficiency, automation, and network economics. It says operators are less likely to rely on uncertain future monetization assumptions as the basis for large-scale investment decisions as data-traffic growth moderates.

Technology and Architecture Direction

Dell’Oro Group characterizes AI-native RAN as increasingly foundational, saying 6G is expected to incorporate AI into the RAN architecture from the start rather than as an add-on after initial deployments. It cites Ericsson’s launch of “ten AI-ready radios featuring in-house silicon with neural network accelerators” as an example of intelligence across RAN layers.

The note says implementation models remain difficult to predict and that industry consensus suggests non-GPU RAN will dominate 6G AI RAN deployments. It also states that open fronthaul discussions are evolving from openness as a multi-vendor strategy toward programmability, automation, virtualization, and software-centric operations.

Fronthaul and Waveform Likelihoods

The document says open fronthaul adoption is expected to be “significant from the start,” while multi-vendor RAN is projected to account for “less than 5 percent of total RAN by 2030.” It links RIC frameworks and software-driven optimization to support AI-native RAN architectures over time.

For waveform evolution, Dell’Oro Group says the most likely 6G waveform outcome is “an evolutionary path built around enhanced OFDM rather than a completely new waveform.” It adds that OTFS remains under evaluation for specific use cases including high mobility and integrated sensing and communications, but it says the alternative has not shown sufficient system-level benefits to replace OFDM at scale.

Use Cases, Traffic Demand, and Adoption Pace

Dell’Oro Group identifies demand-side uncertainty, citing Ericsson’s Mobility Report baseline projections that “total mobile network traffic, including mobile broadband and FWA, will grow 15% to 20% annually over the next five years.” It says these projections assume stable usage patterns driven primarily by existing smartphone behavior and adoption trends.

The note also says ISAC remains a high-activity area in research and standardization, but that ISAC upside is incremental because it “will require more site modifications.” It states ISAC “is clearly not yet the primary driver behind early 6G investment planning.”

Forecast and Outlook on RAN Segments

On adoption timing, the document says uncertainty centers on how aggressively operators will scale 6G and the size of the gap between early adopters and late majority. It characterizes operator expectations that 6G will be more evolutionary, relying more on software and fewer large-scale hardware upgrades, while also noting that competitive differentiation can change once deployment cycles begin.

It assesses likelihood of mass adoption by 2030 as “very likely for 6G and AI RAN, less likely for Cloud RAN, and unlikely for multi-vendor RAN.” It also ties Cloud RAN shifts to a TCO metric, stating operators need “the performance-per-dollar-per-watt gap between custom silicon and COTS plus accelerators” to narrow for a gradual increase in COTS share.

This Analyst Signals brief reflects a neutral, fact-based summary of the original research note.