A perception product, not another aircraft.

Haltere is a product of ABX LIMITED, registered in England and Wales. We are developing event-based motion perception for inspection aircraft, with an initial focus on PX4 multirotors and high-contrast, GPS-denied environments.

The intended role is a reusable software component alongside the customer's autopilot. We are not building a replacement flight controller or asking an OEM to adopt a new airframe.

Start with a problem a customer can fund

Our first customer hypothesis is a drone OEM or robotics integrator with an active industrial-inspection program, a specific perception failure, and an existing baseline to compare against.

The entry offer is scoped, paid feasibility work. We agree the use case, constraints, data access, and acceptance criteria before making a delivery commitment. The initial artifact is a reproducible offline evaluation, not flight-ready autonomy.

Build toward repeatable software revenue

The proposed commercial sequence is paid evaluation, separately scoped reference integration, then OEM deployment licensing and support.

The product test is whether multiple customers can use substantially the same perception core without open-ended custom engineering. Reusable integration, bounded support effort, and demonstrated customer value must support the licensing economics. This is the intended business model, not a claim of existing recurring revenue.

Inspection is the entry point. Broader aircraft and robotics applications would be expansion opportunities only after the initial product earns them.

Why this approach is worth investigating

Event sensors and embedded development ecosystems provide a practical starting point for investigating asynchronous perception. Our technical hypothesis is that local, event-native motion computation can be useful where lighting transitions, motion, and hardware budgets challenge an existing pipeline.

There are established visual-inertial, optical-flow, and bio-inspired alternatives. Event sensing, Rust, and the insect analogy are not sufficient differentiation. The case must rest on measured task performance, total system cost, integration effort, and implementation know-how.

Our technical brief sets out the proposed architecture, limitations, and selected references.

Milestones that would reduce risk

• Technical usefulness: reproducible results against a credible, customer-relevant baseline, including failures and resource requirements.
• Commercial usefulness: a funded evaluation and a customer's decision to pay for the next step.
• Product repeatability: another customer using the same core offer, with measurable integration and support effort.
• Deployment feasibility: one supported hardware configuration and a reference integration that progresses through appropriate safety gates.

These are goals. Specification and architecture are complete; core implementation is starting. No Haltere performance results or flight demonstrations have been published, and no customer or revenue traction is claimed on this site.

The remaining risks include whether the approach beats an appropriate baseline, sensor and compute cost, real-world reliability, integration complexity, and willingness to pay. Capital would need to fund specific evidence-producing milestones, not assume those questions have been answered.

For investors

We welcome conversations with investors who understand early-stage perception, robotics, and embedded software. The useful discussion is about the technical work, customer access, team, and capital required to reach the next milestone.

The website is an introduction, not a substitute for diligence. No funding round, valuation, investor commitment, or tax-relief eligibility is represented here.

The company behind Haltere

ABX LIMITED is registered in England and Wales, company number 17185335. Registered office: 27 Old Gloucester Street, London WC1N 3AX.

For company, team, or partnership enquiries, contact hello@haltere.xyz. For a product evaluation, tell us about your inspection use case.