AI-Native Open RAN Market to Hit $24.5 Billion by 2035: The Investment Case Is Finally Catching Up to the Architecture

The Headline Number

A new market forecast puts the AI-native Open RAN opportunity at USD 3.20 billion in 2025, expanding to USD 24.52 billion by 2035 — a 22.63% compound annual growth rate across the 2026–2035 window. For an industry that has spent the better part of a decade arguing about whether Open RAN would ever scale beyond greenfield curiosity, that is not a rounding error. It is a structural re-rating of the addressable market.

Our read: this is the first forecast of the current cycle that treats AI and Open RAN as a single purchasing category rather than two adjacent stories. That framing matters more than the absolute dollar figure. It reflects what operators are actually buying — intelligent, flexible, interoperable, and cost-efficient radio access network infrastructure — and it aligns with the capex signals we are seeing elsewhere in the market, including a 5% rise in global telecom capex in the first half of 2026 as operators shift spending toward capacity, AI, and network modernization.

Why the Two Stories Are Now One Story

The bull case for Open RAN was always economic: vendor interoperability, reduced dependence on proprietary equipment, and a software-driven cost curve that behaves more like cloud than like radios. The bull case for AI in the RAN was always operational: network optimization, traffic management, resource allocation, predictive maintenance, and — increasingly — autonomous operations.

What the new forecast captures is the convergence of those two curves. Disaggregation without intelligence produces a more open network that is still manually tuned. Intelligence without disaggregation produces a smarter proprietary network. Neither delivers the full prize. The combination — cloud-native network functions supporting scalable, software-driven deployments, with AI/ML sitting across optimization, energy, and assurance — is where the incremental margin actually lives.

That is why the vendor list behind this market reads the way it does: Ericsson, Nokia, Huawei, Samsung, ZTE, NEC, Fujitsu, Mavenir, Rakuten Symphony, Parallel Wireless, Radisys, and Airspan on the RAN and software side; Intel, NVIDIA, Qualcomm, and AMD on silicon; Dell Technologies, HPE, and Cisco on infrastructure; Keysight Technologies on test and validation. This is no longer a startup-only story. It is a full-stack supply chain.

Regional Skew: The US Is Still the Anchor Tenant

The regional detail is where the forecast gets interesting. The US AI-native Open RAN market is projected at USD 1.12 billion in 2025, rising to USD 7.29 billion by 2035 — a 20.66% CAGR, slightly below the global rate but off a much larger base. The US accounted for roughly 91.0% of the North American market in 2025.

That concentration is both an opportunity and a warning. On the opportunity side, it means the US remains the single most important reference market for AI-native RAN procurement, and vendor wins there will set the template for deployments elsewhere. On the warning side, a market that is 91% one country is not yet a global market — it is a beachhead. The global CAGR running ahead of the US CAGR implies the growth differential in the back half of the forecast period comes from outside North America, which is exactly where the multi-vendor, brownfield complexity is highest.

We would flag three drivers behind the US skew: 5G rollout maturity, a diversified vendor base, and sustained spending on AI-native networking. The first is cyclical. The second and third are structural, and they are the ones that matter for a ten-year revenue model.

The Technical Stack Behind the Revenue

Any credible market forecast has to be traceable to a bill of materials. In AI-native Open RAN, four line items do the heavy lifting.

AI-powered RAN automation is the most mature. Traffic management, energy optimization, and network performance tuning are being delivered today, largely through near-real-time control loops. Energy in particular is where the ROI conversation starts for most CFOs, because it converts an architectural argument into an opex line item.

Cloud-native network functions provide the scalability substrate. Without containerized, software-driven deployment, AI models have nowhere to run and no way to scale horizontally across sites. This is the least glamorous layer and the one most likely to be underestimated in revenue models.

Near-RT RIC applications are where the forecast gets its growth premium. The near-real-time RAN Intelligent Controller is the control point for AI-driven network control — the place where third-party and operator-developed applications compete on measurable performance. If the RIC application ecosystem matures as forecast, it becomes a recurring software revenue stream layered on top of infrastructure. If it stalls, the market reverts to being a hardware replacement cycle with better marketing.

Edge AI extends distributed intelligence and low-latency decision-making across the network. This is the piece that connects RAN economics to the broader edge compute build-out, and it is the reason the silicon vendors — Intel, NVIDIA, Qualcomm, AMD — appear on the key player list alongside traditional RAN names.

The Near-RT RIC Is the Swing Factor

If we had to pick one variable that determines whether the 22.63% CAGR is achieved or missed, it is the RIC application economy.

Infrastructure revenue is relatively predictable: it tracks deployment counts and refresh cycles. Application revenue is not. It depends on whether operators can attract developers, whether the interfaces are stable enough to build against, and whether the resulting applications produce measurable, attributable gains that justify their own budget line.

The forecast implicitly assumes the answer is yes. Our view is more conditional. The near-RT RIC has been “about to matter” for several years. What has changed is that AI workloads have given operators a concrete reason to care about real-time control — not as an architectural elegance argument, but as the mechanism that turns a model into a network outcome. That is a materially stronger commercial pull than anything the RIC ecosystem had before.

We would also note the timing coincidence with India Mobile Congress 2026, which opens October 7 in New Delhi with 6G, AI-native telecom, intelligent RAN, edge and cloud computing, and autonomous and zero-touch networks headlining the agenda. When a major regional show organizes its tracks around intelligent RAN and AI-native operations, that is a demand signal, not just a conference theme.

The Vendor Landscape Is Crowded — and That Is the Point

The key player list spans at least twenty companies across RAN software, silicon, infrastructure, and test. For operators, that breadth is the entire thesis: interoperability and reduced dependence on proprietary equipment. For vendors, it is a margin problem. When twenty credible suppliers compete for the same AI-native RAN budget, differentiation has to come from something other than the radio.

The realistic differentiation axes are: silicon efficiency (performance per watt per site), software velocity (how fast new AI applications can be deployed and validated), integration depth (how much of a brownfield network can be upgraded without forklift replacement), and assurance credibility (whether the vendor can prove the AI actually improved the network).

That last point is underappreciated. AI-powered service assurance appears repeatedly in operator priority lists — including HCLTech’s AI-native telecom portfolio announced this week for IMC 2026, which spans autonomous networks, OSS modernization, network monetization and mediation, network cloudification, AI-powered service assurance, Open RAN engineering, intelligent infrastructure, and the evolution toward 6G. When systems integrators start packaging AI-native operations as a services offering, it signals that operators are buying outcomes rather than components — and that the measurement layer is becoming a competitive battleground.

What Could Break the Forecast

Three risks deserve explicit mention.

First, the economics of flat RAN. The total RAN market has been growing in recent quarters, but the AI-native premium has to be funded from somewhere. If operators cannot demonstrate that AI-native Open RAN lowers total cost of ownership — energy, spectrum efficiency, operational headcount — the premium gets negotiated away, and the forecast compresses toward the underlying RAN growth rate.

Second, integration drag. The global CAGR exceeds the US CAGR, which means the growth story depends on markets with more multi-vendor complexity, more legacy, and less standardized procurement. Integration cost is the single most common reason Open RAN business cases fail in brownfield. AI layers on top of that complexity rather than removing it.

Third, the application economy timing. If near-RT RIC applications take longer to monetize than the forecast assumes, the software revenue layer arrives late, and the market becomes infrastructure-weighted — lower multiple, lower growth, same architecture.

The Analyst Take

A 22.63% CAGR to USD 24.52 billion by 2035 is an aggressive but defensible number, provided three things hold: near-RT RIC applications monetize on schedule, edge AI deployments connect RAN intelligence to broader edge compute spending, and operators can prove measurable TCO improvement rather than asserting it.

The more important takeaway is structural. AI-native Open RAN has stopped being a debate about whether disaggregated architectures can work and become a forecast about how much intelligent, interoperable infrastructure the market will absorb. That is a fundamentally better position for the industry than where it stood two years ago.

The US will anchor the early revenue. The growth will come from everywhere else. And the vendors that win will be the ones that can prove — with numbers, not architecture diagrams — that intelligence in the RAN pays for itself. Watch the near-RT RIC application ecosystem. It is the leading indicator for everything else in this model.

Sources

• AI-Native Open RAN Market to Expand Worldwide as AI, 5G, and Open
• Global Telecom Capex Rises 5% in H1 2026 as Operators Shift Spending to Capacity, AI and Network Modernization — TelecomLead
• HCLTech to Unveil AI-Driven Solutions for Telecom at IMC 2026 — ETTelecom
• India Mobile Congress 2026 begins Oct 7 with 6G, AI, cybersecurity in focus — Business Standard