Why Relay
Only a neutral network can aggregate this market
Every incumbent could technically build what Relay is. None of them can actually run it, and the reason is not engineering. It is the single rule that governs this market: nobody feeds a competitor’s network.
Why the seat is empty
A telematics platform that launched a road event network would never see its competitors contribute to it. A carmaker that opened one would attract no other carmaker. The asset is not the technology; it is the position.
This is why Visa is not owned by a bank
Interbank settlement is not run by one of the banks, and card payments are not run by one of the merchants, for exactly the reason above: a shared network whose owner competes with its participants stops being shared. Neutrality here is not a moral posture. It is the asset, and it is the one thing an incumbent cannot acquire.
The densest road network in the world is captive to one app
That is what makes it unbeatable head-on, and it is precisely what makes it possible to route around. No navigation app, no GPS maker and no carmaker will ever contribute to a consumer application they compete with — which makes every one of them a structural candidate for a neutral trunk. We are not building an application to fight theirs. We are building the trunk that lets everyone else rebuild that network together.
Some events only a vehicle can witness
Human-declared reports bring the network to parity — a police check, a pothole, a hazard somebody saw. What creates the asymmetry is the half that arrives with no user involved at all, and it arrives because the federation runs on vehicles rather than on phones.
Suspension travel
Road damage — a pothole nobody reported
Wheel slip
Aquaplaning, on the stretch where it happens
Outside temperature
Black ice, before the first car finds it
Stability correction
A gust on an exposed viaduct
No phone application will ever produce these, and no gantry camera sees them. They reach the network in exactly the same envelope as a driver’s report and are subject to the same quorum — which is why a consumer integrates once and receives both. The event vocabulary is drawn against a navigation app’s interface rather than against a public feed’s schema: public feeds are a source to map into it.
The long tail, which is why signing up does not involve us
Not from one founding partner bringing everything. The app stores are full of navigation, journey and road-community apps with a few hundred to a few thousand users each — individually negligible, collectively the network.
That target imposes a hard technical constraint rather than a convenience: a key struck by hand for every partner is compatible with ten carmakers and not with three hundred small applications. So onboarding is entirely self-service, and the second factor at the moment a key is issued is deliberate — we sell the time it takes to get a key, not the imprudence of handing one out on a session alone.
Three axes, and coverage is not one of them
The commercial competitors are the companies already selling real-time incident feeds by API. They have years of coverage on us and we will not win there for a long time. These are the three where the position, not the effort, decides the outcome.
Reciprocity
Today a carmaker with five million connected vehicles gives its fleet's data to a supplier, then buys back the service built with it. Here what your fleet publishes comes back as every other participant's events on the same wire, because we build no application to keep them in. An incumbent cannot copy that without dismantling its own model.
Independence
Nobody enjoys depending on a supplier who competes with them on the thing they are buying. We build no applications — that is a founding constraint rather than a stage we are at — so there is no product of ours for a partner to be squeezed by later.
Developer experience
This market is sold in long contracts at large amounts. Nobody serves the developer who wants a hundred requests, a key without a sales call, and a sandbox that is not empty. Sign in with an emailed code, create an organisation, issue your own key behind a second factor, and read live national traffic from the first request.
Neutrality is a constraint before it is an argument
A position only works while it costs something. These are the three things it costs us, written down where a partner can hold us to them.
We do not build applications
No consumer app, no navigation product, no fleet dashboard. A partner is never one roadmap decision away from competing with their supplier.
The wire carries no publisher identity
Not a name, and not a stable pseudonym either — a forever-stable pseudonym is a fleet-coverage map missing only its label. We can attribute an event internally, for trust and for billing; nobody on the network can.
Publishing is free, permanently
Friction on the publish path starves the network, and a network that starves is worth nothing to the people paying to read it.