The 5G correction is over, but growth is confined to AI RAN
In sum – what we know:
- Flat to 2030 – Dell’Oro’s base case has worldwide RAN revenue essentially flat in nominal dollars through the forecast period, with no sustained expansion.
- A narrow growth set – AI RAN, massive MIMO and cloud RAN grow within a stable top line, but only private wireless brings genuinely new enterprise capital.
- 6G stays a question – Dell’Oro’s base case is that 6G is “just another G,” though slowing traffic growth and machine-driven demand could change that math.
The radio access network market will stay essentially flat through 2030, but there will still be a real shift in what operators are buying, according to Dell’Oro Group. Stefan Pongratz of Dell’Oro opened the Intelligent RAN Forum with a two-part message: the post-5G inventory correction is over, and the growth that remains is concentrated in a handful of segments rather than spread across the market.
Dell’Oro sizes the RAN market at roughly $30 billion to $40 billion a year in manufacturer revenue — made up of hardware, software, baseband, and radio, and excluding services and passive antennas.
After a couple of strong years, RAN revenue fell about $9 billion between 2021 and 2024 and was flat in 2025. Dell’Oro has now logged three consecutive quarters of year-over-year growth, which Pongratz said confirms the shift from 5G coverage buildout to capacity — “it’s really a transition from coverage to capacity that is now largely behind us.”
But that does not mean spending takes off again. Worldwide RAN revenue is expected to be flat this year, and Dell’Oro’s base case through 2030 is the same.
What’s behind the decline
The reason for the decline, Pongratz argued, is that 5G delivers more capacity per dollar than 4G did — roughly a 2x boost — while mobile data traffic is growing only 20% to 30% a year. That gives operators an opportunity “to slow things down for a bit,” he said. “You do get more bang for your buck” with 5G.
The trough is visible across regions. North America and Japan fell about 50% peak-to-trough, and India dropped roughly two-thirds from 2023 to 2025. In the first half of this year, EMEA held up because those markets are earlier in their 5G rollouts, while China — still the largest market — pulled back, and Asia-Pacific outside China improved.
Pongratz was also clear that the traditional market still calls the shots. For all the attention on AI RAN, cloud RAN, private 5G, fixed wireless access and small cells, the spending envelope is still dominated by 5G sub-6 GHz macro RAN and mobile broadband.
Beneath that stable top line, the composition is shifting. Dell’Oro expects growth between 2025 and 2030 in AI RAN, massive MIMO as networks push into higher bands, and cloud RAN, which has had a slow start. AI RAN, Open RAN and 6G are each forecast to become a significant share of total RAN revenue by 2030.
The catch is that almost none of this grows the overall market. Most of it is changes to product mix, architecture and configuration — a cloud RAN deployment that replaces a traditional one redirects spending rather than adding it. The one exception, Pongratz said, is private wireless, which brings genuinely new capital from enterprises.
The 6G question
Dell’Oro’s base case for 6G is that it is “just another G.” Generational transitions have not reversed the flat revenue trajectory through 4G and 5G, and Pongratz does not expect 6G to do it on its own. But AI changes the odds. “The likelihood that 6G will be more than another G is greater now than ever before,” he said.
The reasoning rests on traffic. 5G coverage has not yet reached 60%, so a lot of 5G base stations are still to be deployed — but year-over-year comparisons get harder from here. On the capacity side, data traffic growth has slowed to roughly 2.2x over the period, or about 17% a year — unusual in a market that has always been short on capacity rather than sitting on it — Pongratz, who has covered RAN since 2010, called it “uncharted territory.” The variable that could change the math is the mix of humans, AI agents and machines generating that traffic. More machine-driven demand could justify new spectrum and pull 6G forward; without it, 6G timing could stretch for most operators.
One structural constraint limits how far any of this can move the top line. RAN is about a fifth of total wireless capex, and the two have held a 98% correlation. “If you want to change the capex trajectory, you have to change the revenue trajectory,” Pongratz said.
On the forum’s central theme, Pongratz was direct. Intelligent RAN, as Dell’Oro defines it, is the part of the broader AI RAN opportunity focused on applying AI and automation to the RAN itself — spanning operations, maintenance and the services layer, not just spectral efficiency.
And it does not have to mean GPU RAN. Today’s adopters are pragmatic, he said: mostly single-purpose, non-GPU deployments on distributed 5G architectures that lean on the existing install base to improve performance, energy efficiency and operations.
Dell’Oro projects about $35 billion in cumulative AI RAN revenue over the next five years, with GPU RAN crossing $1 billion by the end of the forecast period. Those are real numbers, but not transformative ones. Pongratz was explicit that AI RAN and Intelligent RAN are not expected to materially expand the overall RAN market. The point is to make the RAN better — cheaper and less complex to run — as networks get more complicated heading into 6G and higher frequencies.
What it means for vendors
Pongratz closed with the stakes for the companies selling this equipment. The total telecom equipment market he covers runs about $100 billion a year, and the top five vendors hold 96% of the RAN portion.
His guidance: expect a flat market that is not static, with growth pockets in AI RAN and 6G, and a market that is becoming more polarized, concentrated and geographically divided. Vendors have to build portfolios that work whether human-centric traffic patterns persist or AI and machines reshape demand. In Dell’Oro’s framing, Intelligent RAN and AI RAN are the bridge between those two outcomes — improving the economics of today’s networks while laying groundwork for whatever comes next.
Table of Contents
