The missing sense of physical AI — beyond GPS, lidar & cameras
Every autonomous machine — humanoid, forklift, drone — must know one thing before it moves: where it is. And with it: where every human stands, where every other machine is, and what the environment itself is doing — in real time. Where GPS dies and cameras go blind, we sell the sense that does not fail.
* Customer-reported payback across deployed reference sites — earned before a single robot; every site is robot-ready. † Versus wired alternatives delivering both positioning and sensor data.
The big bang of physical AI
“1.3 billion AI robots by 2035 — 4 billion by 2050.”
Humanoids, vehicles, vessels and drones, working hand in hand with people across industry and society. Every single one of them must answer the same question before it moves: where am I, where is everyone else, and what is happening around me? Language models gave machines reasoning. They did not give them a place to stand. That missing sense is spatial intelligence — and we have spent 14 years building it, for the environments where the work is actually done.
Everyone solves a piece. Nobody owns the layer.
AI cannot optimize what it cannot see. Every autonomy stack is blind somewhere — the layer is where the blindness ends.
Through the eyes of the humanoid
One source of spatial truth across the entire site: an autonomous 3D-positioning sensor mesh carrying real-time positioning, environmental awareness and shared spatial context in one unified platform. Humans and machines share the same workspace — and the same reality. A near-miss that never happens.
Survived, not simulated
Every figure below is pre-robot ROI — paid back by tracking and sensing alone, before a single robot rolled in. And every one of these sites is robot-ready today: the sense layer is already in the walls when the fleet arrives. That is the reason to invest now — not when robots are proven and everyone wants the same floor.
“No other technology would have worked in our difficult steel and concrete-filled environment; it was the only viable choice.”
Machine-level energy intelligence across manufacturing — ~60% lower upfront investment than alternatives.
Ladle tracking and pre-heating in furnace zones beyond 150°C — where competing technologies proved ineffective.
Wheelchair and asset tracking across four terminals, four levels, ~70 gates — one million+ sq ft.
225,000 m² of steel works — ±0.5 m precision in 18 zones, ±5 m across 95% of the site, at 90–95% lower network cost than cabled alternatives.
ROI figures customer-reported per deployment — all earned pre-robot; reference sites are robot-ready.
The hardware of the layer
A compact product family — Sense360 tags, positioning beacons, sub-metre UWB click-ons and hybrid gateways — deployed without wired infrastructure, carrying 3D positioning and live sensor data on the same network. The unified adaptive 3D sense mesh, available out of the box.
One gateway unifies the radios — mesh positioning over sub-GHz, UWB, WiSUN, LoRa and mioty — and third-party industrial sensors join that same mesh directly or through the IO click-on, with thousands already compatible. That is what makes it one network instead of five. Most sites are live within hours.
The platform
Any world model. Any silicon. Any robot. One layer for autonomous systems of every kind — humanoids, AMR fleets, vehicles, drones. A patented hybrid wireless network through a single unified gateway — positioning that scales from site-wide ±5 m coverage to ±0.5 m 3D precision by deployment. TPM-secured. Cloud or on-premises. Integrates with existing industrial software and digital twin environments. Developed and produced in Europe, operated at a 99.9% SLA with multitenancy and RESTful APIs — Security by Design aligned to the EU Cyber Resilience Act, with encrypted, authenticated communications and signed over-the-air updates.
Technology partnerships & ecosystems
Member of Siemens Xcelerator, the open digital business platform. Patented technology relationships and ecosystem integrations with global leaders — the layer is vendor-neutral by design and rolls out alongside the automation, edge and fleet software already running on your site.
Roadmap · coming
Three tracks converge into one layer — from a component you buy to the standard every autonomous fleet is built on. Early-access programs are open: reserve a slot and build with us.
Every world model reads the layer natively — zero integration per robot.
Every new robot ships swarm-ready — the layer scales with the fleet.
One installation serves every fleet — the layer becomes the default.
Roadmap items are in development — dates and specifications in the data room. OEM licensing of the patented positioning technology: contact@wittra.se.
Beyond the factory floor
Wittra creates controlled environments anywhere — physical AI is no longer restricted to factory floors. From the manufacturing line to the warehouse aisle, from the mine face to the airport apron, to the perimeter fence — where the work is dark, dusty, underground or GPS-denied, the layer holds. And beyond: construction, agriculture, commercial prop-tech and smart-city operations run on the same layer.
Start today and join the big bang in Physical AI
A proof-of-value on your site replaces the slow, expensive pilots typical of the category — deployed without wired infrastructure, at 70–80% lower total cost of ownership than wired alternatives. Your site, your assets, your numbers — and a quote shaped to your needs.
Partnerships: partner with Wittra · Investors: reach the founders