Ericsson says mobile networks can track drones, with commercial sensing pilots due in 2027

The pitch is simple: the antennas are already there, pointed at the ground. Ericsson says its Integrated Sensing and Communication (ISAC) technology turns existing mobile networks into a distributed sensor array. That array can spot and follow small drones, according to The Register, which published the details on 2 November.

The rollout is staged. Small-scale, sensing-only pilots with selected customers start in 2027, the report says. Broader deployments follow from 2028. No operator names, no pricing and no coverage figures were given.

The mechanics matter more than the marketing. When one cell site transmits, several receiving sites pick up the reflection off a drone. The network then works out position and track from those multiple perspectives. Ericsson describes the job as complementing radar and other sensors rather than replacing them. It says ISAC needs sensing-capable radio hardware, dedicated compute and software. The Register notes it follows tests over the past few months.

Ian Fogg, research director at CCS Insight, told The Register that sensing has other uses. These include tracking boats in a port, traffic patterns or weather. But drone detection in civilian areas is the most pressing need that other technology cannot easily solve, he said. He argued mobile networks are well placed because their antennas already cover low elevations close to ground level, unlike traditional radar. Fogg added that a standards-based 6G method could improve precision by letting different suppliers' base stations interoperate. New 6G spectrum should also make sensing easier to deploy on wide channels, he said.

The military got there first

Defence interest is not hypothetical. Tom's Hardware reported on 2 November that NATO ran a Digital Backbone Experimentation exercise, DiBaX 2026, in Latvia between 19 and 30 October. The exercise tested whether 5G towers could passively detect and track flying objects, specifically drones. The publication describes the solution as AI-agent-assisted. It says the exercise examined whether large telecom networks could take on a second role in national security. Multiple networks and satellites would add resilience to disruption.

The same piece notes a second demonstration in the United States. AT&T and Ericsson showed drone detection using advanced network sensing over a 5G network near a large sporting venue. That is the part worth watching for anyone thinking about private 5G. The sensing function rides on radio and compute that enterprises already buy for coverage, not on a separate radar product.

Tom's Hardware also raises the obvious objection. If cell masts become sensors, they become targets. The article says the authors would worry about 5G masts becoming an even more attractive target. Ericsson's material does not address that point.

"Mobile networks are useful as drone detection solutions because their antennas are already covering low elevations, close to ground level where drones often fly, unlike traditional radar," Fogg told The Register.

Demand is easy to explain. The Register notes that small, low-cost drones have closed airports on safety grounds. The latest incident was at Jersey Airport over the Halloween weekend. Activity has been rising near military bases and other sensitive sites. In the UK, military personnel have been granted authority to neutralize unwanted flying intruders, the report adds. Scaling detection across large areas is the hard part. That is where a network of towers beats a handful of dedicated sensors.

What it means for private 5G buyers

For industrial and campus networks, the practical question is whether sensing arrives as a software feature on hardware sites already own. Ericsson's answer, as reported, is that ISAC requires sensing-capable radios and dedicated compute. So it is not a firmware switch on any 5G installation. That fits the wider pattern: private 5G vendors are pushing capability into the radio and edge compute layers. Buyers are being asked to plan for functions beyond connectivity.

Context from the chip supply chain suggests the compute side is tightening. Semiconductor Engineering's 30 October week-in-review reported SK hynix expecting memory supply constraints through 2030. DRAM and HBM volumes are effectively sold out through 2029, and Samsung says its 2027 HBM supply is already sold out. The same roundup noted Anthropic warning in its IPO filing that HBM shortages and limited TSMC capacity could constrain AI expansion. Edge sensing workloads are small next to model training. But the shortage sets the tone for anything that needs dedicated accelerators.

Other recent work points the same direction. Semiconductor Engineering's 3 November paper roundup lists IBM Research work on meter-scale quantum interconnects between large quantum processors. It also lists KU Leuven and EURECOM research on secure data memory-dependent prefetching. That is a reminder that hardware security is becoming a design constraint rather than an afterthought. None of it is drone-specific, but it sketches the stack that network sensing would sit on.

There are cautionary notes elsewhere. The Hacker News reported on 1 November that a CVSS 10.0 flaw in Canonical's LXD and MicroCloud, tracked as CVE-2026-112816, let an attacker log in as a trusted client without that client's private key. The attacker needed network access to the HTTPS API and the trusted client's public certificate. Canonical has shipped fixed snap versions, the report says. A CISA assessment dated 30 October found no known exploitation. The lesson travels: the more functions a network hosts, the more the API surface matters.

The same outlet reported on 31 October that Cloudflare mitigated 18.2 million network-layer DDoS attacks in the third quarter of 2026. It also absorbed 14.4 trillion HTTP DDoS requests, or about 8,226 attacks per hour. Short-burst attacks under 10 minutes and 1 Gbps rose 208%. Media, IT services, gambling, software and telecom were the most targeted sectors. Networks that double as sensors inherit that traffic profile.

Timing and the standards question

Ericsson's 2027 pilots land before any 6G standard exists. That is why the company frames the first phase as sensing-only and small-scale. Fogg's point about interoperability is the one to hold onto. A single vendor's base stations sensing on their own is a different product from a multi-vendor network where reflections are pooled. The latter needs standards, and standards need time.

Competing timelines in adjacent markets show how slow this can be. Light Reading reported on 30 October that the FCC moved to knock down copper retirement roadblocks in California. That is a reminder that even retiring legacy infrastructure takes regulatory years. On 29 October, Light Reading also reported Lumen announcing three initiatives aimed at networking for AI and quantum eras. It warned that AI traffic growth could strain India's networks as telcos struggle to monetize capacity. Sensing is one more function to fund.

For now, the only dated commitments are Ericsson's. Pilots come in 2027, with wider deployment from 2028. The technology is hardware-dependent and sold as an addition to radar rather than a substitute. Whether operators bite will depend on two things. Regulators and defence buyers must treat coverage as a security asset. And the compute must be available at the edge when the reflections come back.