For most of a decade, the telecom industry looked for the enterprise buyer who would finally pay a premium for what 5G could do. We pitched smart factories. We pitched connected hospitals, autonomous fleets, and immersive venues. The networks got built. The premium never came.
That buyer existed the whole time. It was simply wearing a uniform.
Téral Research's new report on 5G for defense and military use cases finds a market that has quietly assembled itself out of exactly the capabilities the commercial world declined to pay for: deterministic latency, guaranteed isolation, sovereign supply chains, and radios that do more than move bits. It is small today and growing at a rate the commercial RAN market has not seen since the early 4G build. And nearly everyone modeling the wireless infrastructure business is still leaving it out of the forecast.
The scale problem, stated plainly
Global defense and military spending reached $2.7 trillion in 2025. The top ten spenders, in order, were the United States, China, Russia, Germany, India, the United Kingdom, Ukraine, Saudi Arabia, France, and Japan.
Hold that number against the industry most of our readers work in. Global telecom services revenue in 2025 was $1.8 trillion. Global telecom capital expenditure was $290 billion.
Defense spending is now larger than the entire global telecom services business, and it is growing while telecom capex contracts. Within that $2.7 trillion, we estimate that spending touching 5G accounted for at least 1%, or roughly $27 billion.
One percent. That is the number to sit with. It is not a mature allocation, it is a first tranche. Every structural pressure we can identify points in one direction, and we are only looking at the tip of a very large iceberg.
Two forces, one conclusion
Two things happened at once, and the industry has been slow to connect them.
The first is that geopolitical risk stopped respecting borders. Undersea cables, satellite ground stations, port logistics, power grids, and the networks themselves are now contested infrastructure. The distinction between a military network and a critical national network has eroded to the point where it is no longer a useful planning assumption.
The second is Ukraine. The war has restructured the operational logic of modern warfare around cheap autonomous systems, real-time sensor fusion, and communications that must be reconstituted in hours rather than months. Fixed, proprietary, long-procurement-cycle communications architecture is a liability in that environment. Software-defined, rapidly deployable, commercially sourced architecture is an advantage.
Put those two together and you get a defense buyer who wants precisely what 3GPP has spent ten years standardizing.
What 5G actually brings to the fight
The attribute list is familiar to everyone reading this: eMBB, URLLC, mMTC, network slicing, millimeter wave. What has changed is who values them and why. Three shifts matter most.
The base station became a sensor
This is the development the commercial industry has underestimated. Rather than treating communication and radar as two separate systems, 5G-Advanced and 6G standards let the same radio waves that carry data passively detect, track, and map physical objects in the environment.
The implication is that a dense 5G deployment is also a distributed, dual-purpose radar system. In a threat environment defined by small, low, slow, inexpensive aerial systems, that is not an incremental feature. It is a counter-drone layer built on infrastructure that is already being funded for another reason.
Note what this does to the economics. Sensing is not a new capex line so much as a second return on capex already committed. For a defense budget owner, that is the most persuasive kind of argument there is.
Open RAN answers questions the commercial market never asked urgently enough
Open RAN has spent years being justified on total cost of ownership, and the argument has always been contested. Recast it for a defense buyer and it stops being contested at all.
Breaking open proprietary, closed telecom architectures moves network intelligence into software. For military use, that architectural shift addresses four vulnerabilities directly: supply chain provenance, vendor lock-in, electronic warfare resilience, and the ability to deploy and reconfigure in the field at speed. A network you can inspect, modify, and re-host on domestic hardware is a different asset from one you lease as a black box.
The U.S. Department of War's FutureG office is now pushing this past interfaces to the stack itself. OCUDU, the Open Centralized Unit / Distributed Unit project, aims at an open-source CU/DU software stack for 5G and 6G, developed with the Linux Foundation and backed by an unusually broad vendor coalition. FutureG's Thomas Rondeau has framed the ambition in explicitly historical terms: "what Linux did for internet infrastructure, and what Kubernetes did for cloud, OCUDU is meant to do for the radio access network."
We would offer a friendly caution. The industry has made a version of this promise before, in 2012 with NFV and again with the first wave of Open RAN. Both under-delivered, and not for technical reasons. Integration cost migrates from the vendor to the buyer, and the buyer needs a software organization capable of absorbing it. What is different here is the buyer. A defense department has both the motive and the budget to build that organization, and it is not optimizing for quarterly capital intensity.
LEO became the tactical backbone
Non-terrestrial networks in a defense context are not a coverage story. Low Earth orbit constellations offer the latency and throughput profile that data-intensive tactical operations require, without dependence on terrestrial infrastructure that can be destroyed, jammed, or simply never existed in the operating area.
This is why spaceborne open gNBs matter more than their current market size suggests. The architecture being validated now is the one that gets scaled in the 6G cycle.
The forecast
Téral Research sizes the addressable market across five categories: 5G slicing, 5G sensing, private 5G including network-in-a-box, 5G mmWave for tactical bubbles and kits, and 5G NTN spaceborne open gNBs.
That market reaches $1.8 billion by year-end 2026 and grows to $11.3 billion, a 61% five-year CAGR.
A few things to draw out of that.
Slicing leads, and the gap widens. 5G slicing accounts for 34% of the market in 2026, rising to 37% in 2031. Slicing is the critical nerve of the battlefield: it is what allows a single physical network to carry command traffic, sensor feeds, logistics, and coalition partners with guaranteed isolation between them. Sensing follows, then private 5G, then mmWave, then NTN.
It is worth pausing on the irony. Network slicing was 5G's flagship enterprise feature, and commercially it has been the hardest to sell. Enterprises wanted the outcome and would not pay for the mechanism. The defense buyer wants the mechanism, understands exactly why it is hard, and will pay for it.
The U.S. dominates and stays dominant. As the world's largest defense spender, the United States leads throughout the forecast period, followed by Europe, the Middle East, and Asia Pacific. European rearmament is the variable most likely to move this mix, and we are watching it closely.
A 61% CAGR is not a large market, it is a fast one. At $11.3 billion in 2031, this remains a fraction of a roughly $35 billion commercial RAN market. But commercial RAN is forecast at roughly flat over the same window. For vendors deciding where to allocate scarce R&D during the 5G-to-6G lull, growth rate is the relevant variable, not absolute size.
Téral Research's View
We wrote in June that 5G's infrastructure was solid and its monetization layer was what failed. This report is the most concrete evidence yet that the failure was one of customer identification rather than of technology.
Four points for the operators, vendors, and investors reading this.
One. If you build wireless infrastructure and you do not have a defense strategy, you have a gap in your 2031 revenue plan. The procurement cycles are long enough that positions taken now determine who is in the room in 2029.
Two. Sensing is the segment to watch, not slicing. Slicing is the largest category and will stay largest, but sensing is where the standards work in 5G-Advanced and 6G converts most directly into a capability that has no commercial substitute and no easy alternative supplier.
Three. OCUDU is a strategic event, not a research project. An open-source RAN stack with government backing changes the competitive floor for every incumbent. It also gives Open RAN something it has lacked: an anchor customer that values sovereignty over price.
Four. Dual-use cuts both ways. Defense requirements are about to start shaping the commercial roadmap, in sensing, in resilience, in supply chain provenance, and in how 6G handles trust. The industry spent 5G trying to sell verticals on telecom's priorities. In 6G, one vertical is going to set them.
We have watched the telecom industry look for its buyer for ten years. It turns out the buyer was never going to be found in a factory pitch deck. It was in a budget line that grew to $2.7 trillion while we were looking somewhere else.
The full report, "5G for Defense and Military Use Cases," including five-year forecasts by category and region, is available to Téral Research clients. For access or a briefing, contact us at contact@teralresearch.com.