The Operators Will Build the Sensing Network. Here's How.

Sebastian Barros published a piece this week that every telecom executive should read twice. His central argument is precise: the mobile network is the only infrastructure in America with the geographic footprint to become a nationwide sensing grid, the physics of ISAC make it technically feasible today, and yet no operator has built it. Hidden Level, a passive RF startup from Syracuse, got there first.

He's right on all three counts. But the story doesn't end with Hidden Level winning by default. It ends with operators deciding whether to build the software layer that turns their existing hardware into a sensing platform — or cede that value permanently to companies that will.

The Infrastructure Argument Is Settled

Barros cites the Wireless Infrastructure Association's March 2026 count: roughly 255,000 macrocell sites and 198,000 outdoor small cells. That's the addressable grid. No startup, no defense contractor, no government agency can replicate it. It took forty years and hundreds of billions of dollars to build, and it's already running 24 hours a day.

The Massive MIMO radios deployed for 5G changed the physics equation. These aren't the single-antenna base stations of the LTE era. They're phased arrays with 64, 128, sometimes 256 antenna elements, each generating a beam that can be steered in real time. When you have that kind of spatial resolution, the reflections that used to be filtered out as noise become signal. You can read them. You can track what caused them.

The object being tracked doesn't need to transmit anything. That's the sentence that matters for drone detection, for the 600 incursions the FBI logged at World Cup host cities, for the scenario Barros opens with at SoFi Stadium in 2028. Military-grade drones that transmit nothing are invisible to every system built to listen for RF emissions. They are not invisible to a network reading reflections of its own signals.

Why Operators Haven't Moved

Barros frames this as incumbents being slow. That's partially true, but the more precise explanation is structural.

Carriers are licensed to carry traffic. Their regulatory frameworks, their network operations centers, their SLAs, their vendor relationships — all of it is organized around throughput, latency, and uptime. Sensing is a different product with a different buyer, a different sales cycle, and a different liability profile. A carrier selling capacity to a transit authority is a familiar transaction. A carrier selling situational awareness to a federal security agency is not.

The July 1 law change Barros references matters precisely because it creates a defined legal framework for the second transaction. Before July 1, an operator offering sensing-as-a-service to a non-telecom buyer was navigating genuine regulatory ambiguity. That ambiguity is now resolved. The question shifts from "can we do this?" to "how do we operationalize it?"

That's a software problem.

What the Software Layer Looks Like

Hidden Level built a passive RF sensing network by deploying its own receivers. That's one approach. It works, and they deserve credit for executing it. But it doesn't scale to 255,000 sites without a capital program that rivals building a new carrier from scratch.

The alternative is to activate the sensing capability that already exists in deployed 5G hardware — without replacing it, without a new rollout, without a decade-long infrastructure program.

This is what ISAC software platforms do. The base station hardware is already there. The Massive MIMO arrays are already there. The fiber backhaul is already there. What's missing is the software stack that:

• Extracts the sensing signal from the communication waveform in real time
• Fuses returns across multiple base stations into a coherent 3D picture
• Runs inference at the edge to classify objects — drone, aircraft, vehicle, person — without sending raw RF data to a central cloud
• Exposes the output through an API that a transit authority, a stadium operator, or a federal agency can consume

That stack runs as an xApp on the O-RAN RIC. It doesn't require a forklift upgrade. It doesn't require new spectrum. It requires software deployed to existing infrastructure — and it can be turned on across thousands of sites in the time it takes Hidden Level to deploy a few hundred dedicated sensors.

The Race Barros Describes Is Already Running

Barros writes that "the race ends before 6G begins." We agree. The window is the next 24 to 36 months, while 5G Massive MIMO deployments are still being completed and before 6G standardization locks in new architectural assumptions.

The operators who move in this window will own the sensing layer. The ones who don't will find that a combination of startups, defense primes, and hyperscalers have built the value stack on top of their infrastructure — and are capturing the margin that should have been theirs.

The infrastructure argument is settled. The physics argument is settled. The legal argument was settled on July 1.

What remains is execution: deploying the software that activates what's already standing.

---

OranSense builds the AI sensing and inference platform for 5G and 6G networks, powered by NVIDIA Aerial SDK and O-RAN open interfaces. The InfraSense platform is available for operator evaluation today.