Five years of vendor advocacy, government subsidies and operator commitments have moved Open RAN from 21% to 22% of operator adoption, according to GSMA Intelligence data cited in a new Observer Research Foundation analysis.

Dell'Oro Group's February 2026 Open RAN Report now projects multi-vendor RAN at less than 5% of total RAN by 2030, a downward revision from the firm's earlier 5–10% range. The forecast caps a year of scaled-back rollouts, delayed launches and one greenfield poster child — EchoStar — exiting the carrier business in September 2025.

Where do the live deployments actually stand?

The O-RAN Alliance and the Telecom Infra Project counted 45 ongoing Open RAN deployments and trials across 27 countries and 31 operators by mid-2025. Concrete milestones have landed:

• AT&T completed the first commercial call on its Open RAN network using a Fujitsu 1-Finity radio connected to an Ericsson distributed unit.
• Bharti Airtel deployed 2,500 Open RAN sites in rural India with Mavenir.
• Rakuten Mobile runs a fully virtualised, cloud-native Open RAN footprint covering 98% of Japan's population.
• AT&T signed a US$14 billion deal with Ericsson — the largest operator commitment to date.

Yet the headline metric barely moved. By late 2025, 22% of operators were in the initial phases of deploying open networking technologies, against 21% in 2019. Omdia's 2026 Open RAN Operator Survey put meaningful O-RAN architecture adoption at roughly 40%, with 13% actively deploying in production and 27% using O-RAN principles to guide evolution. GSMA Intelligence reported Open RAN's status as a technology priority was nearing an all-time low.

Why has adoption flatlined?

Performance remains the predominant barrier in operator surveys. Multi-vendor interoperability, the architectural premise of Open RAN, proved harder in practice than the standards implied. Operators found that carrier-grade performance in multi-vendor configurations required integration work no vendor's sales cycle had priced in. Every additional supplier introduced further testing requirements and exposed accountability gaps.

Cost savings proved equally elusive. Integration and testing costs frequently offset procurement savings, and dense urban macro networks remain an uneconomic use case. Smaller vendors came under structural pressure: contracts increasingly go to Nokia, Ericsson, Samsung, NEC and Fujitsu — the incumbents with the capital to absorb integration complexity.

Two exits compounded the pressure. Mavenir, one of the technology's most visible advocates, dropped its radio hardware business in 2025 to focus on AI software and now carries roughly US$1 billion in debt. EchoStar's September 2025 closure of its Dish 5G network — long held up as a commercial proof-of-concept — removed a marquee reference customer.

Where does the momentum actually sit now?

Geopolitics has kept policy intact. The Quad — the US, Australia, India and Japan — committed US$20 million in 2026 to the first commercial Open RAN deployment in the Pacific, with Palau's network expected to be operational by early 2027. The package, built on a 2023 pilot, was framed explicitly as a counterweight to Huawei after Vodafone's 2019 discovery of hidden backdoors in Huawei equipment deployed in Italy.

The bigger shift is convergence with AI. The AI-RAN Alliance, formed in February 2024, now coordinates the move from rule-based RAN optimisation to machine-learning models capable of running telecom and AI inference workloads on the same base-station hardware. In October 2025, Nokia and NVIDIA announced a US$1 billion strategic partnership to run Nokia's AnyRAN software on NVIDIA GPU-accelerated platforms. Early 2026 functional tests with T-Mobile US and SoftBank demonstrated concurrent video streaming, AI queries and live 5G operations on shared infrastructure. In March 2026, Ericsson and Nokia announced a joint initiative toward autonomous networks — an unusually explicit sign that the centre of gravity has moved.

What is the realistic path from here?

The ORF analysis concludes: "Open RAN is best understood as a selective deployment strategy rather than a universal replacement for traditional RAN, performing best in greenfield networks, rural coverage, strategic diversification and geopolitically sensitive deployments — as illustrated by the Quad's initiative in Palau."

The market still carries a 39.4% CAGR forecast, off a low base. As the industry positions for a US$100 billion 6G cycle, open interfaces now look more useful as substrate for AI-RAN than as a standalone commercial thesis, with Quad-backed deployments in the Pacific likely to remain the clearest proof points through 2027.