A quiet change in pest control could become very difficult to ignore

The Pentagon has a GPS problem. Modern electronic warfare can spoof or jam satellite signals in seconds, turning a sophisticated armed drone into a $50,000 paperweight. The military’s answer is to remove the GPS dependency entirely, and that means embedding navigation intelligence directly into the chip, not in a data link back to a ground station.

That is exactly the market BrainChip Holdings (OTCQX: BRCHF; ASX: BRN) is chasing right now. And the evidence that real money is arriving is specific enough to be worth understanding.

What Neuromorphic Actually Means on a Drone

A conventional drone autopilot runs a continuous loop: sample every sensor, crunch every pixel, update every actuator, repeat. That cycle burns power whether anything is happening or not. Neuromorphic semiconductors are designed to process data in a similar way to a human brain. The machines react only when something happens, reducing energy consumption dramatically in the right workloads.

Small tactical UAVs face a fundamental constraint: batteries powerful enough to sustain flight leave minimal energy headroom for onboard computation. Traditional deep neural networks running on graphics processing units consume power budgets that are often incompatible with platforms weighing under two kilograms. This energy bottleneck has pushed military drone developers toward neuromorphic processors that mimic sparse, event-driven computation.

The payoff is a drone that can navigate a contested urban corridor, classify targets, and avoid obstacles without a constant stream of GPS data, all on a battery that also powers the rotors.

The Contracts Are Real. So Are the Complications.

BrainChip Holdings, with a market capitalization of about US$284 million as of October 12, 2026, has disclosed work tied to a US$1.8 million Air Force Research Laboratory (AFRL) SBIR contract focused on mapping complex RF and radar signal processing algorithms onto neuromorphic chips. BrainChip has said Raytheon is the major defense contractor supporting the completion of that AFRL project, and the work has referenced micro-Doppler style activity classification.

Raytheon’s role here is worth holding in context. Raytheon and Lockheed carry record backlogs in precision munitions, yet production capacity has not yet caught up with the order pace.

The project intends to validate a new capability to embed sophisticated radar processing solutions in SWaP-C constrained platforms utilized in a wide variety of systems, including threat detection, air defense, and guidance use cases in mobile platforms such as drones, robots, aircraft, satellites, and projectile-based defense systems.

The sales funnel is strong and growing quarter-over-quarter, with significant traction in defense and wearables. The company has secured key partnerships and deployments, such as with Raytheon, Onsor, and Parsons, and it has said it has sold through 25% of Akida 1500 inventory.

That looks encouraging until you read the balance sheet alongside it.

The Math That Matters

BrainChip booked about US$1.223 million of revenue in the first half of 2026 while reporting net cash used in operating activities of about US$11.25 million over the same half-year period. As of June 30, 2026, it reported roughly US$31.0 million in cash and cash equivalents.

The first production batch of its AKD1500 chip arrived in 2026, about 2,000 units, and management has discussed yields coming in lower than expected. A licensing model could eventually make the unit economics work. Interest is surfacing in several corners: space applications, newer offerings like Akida Pico, and defense.

Defense primes converting development contracts into volume orders is the mechanism that determines whether small-chip suppliers ever see real revenue. Lockheed’s record $230 billion backlog shows how long that conversion cycle runs even for established suppliers.

Risks Worth Naming

The gap between about $1.2 million in first-half revenue and roughly $11 million in first-half operating cash burn is not a rounding error. A dilutive capital raise before the next major product milestone is a real possibility. Defense procurement cycles are long, and a $1.8 million AFRL contract funds development, not volume orders. Conventional AI chips are also getting more power-efficient each generation, compressing the window where neuromorphic efficiency alone justifies the engineering complexity.

The Bigger Picture

GPS-denied autonomous flight is not a niche requirement. It is fast becoming the baseline spec for any tactical drone operating in a peer or near-peer threat environment. A drone, a satellite, or a soldier’s sensor has a fixed battery, often no network connection, and a need to react in milliseconds. Those are exactly the conditions where brain-like chips can have an edge.

BrainChip may be one of the few publicly traded companies that is primarily identified with neuromorphic, spiking neural network chips. Whether it can survive long enough to collect volume contracts is a different question, and right now the answer depends on whether its next product cycle lands on schedule and whether a defense prime converts a development deal into a production order. Worth watching closely. Not yet a comfortable hold.