Software for the
physical world.
Nextwrks is a Dubai practice that builds intelligence for roads, buildings and machines. What follows is the work: what each system is, where it stands today, and who we want to build it with.
We build the intelligence that roads, buildings and machines run on.
Nextwrks is one practice in Dubai with four divisions: Synapse, TENET.WORKS, ZARC and Nextlab. They work as a single team, from the first sketch to the deployed system.
Visit nextwrks.dev →A few projects, each finished completely.
We take work where failure is expensive: governments, secure facilities, mobility operators and autonomous systems. One practice carries each project from strategy to deployment, to one standard.
Visit nextwrks.dev →ZnakOS is an early-warning system for roads.
When a car stops on a fast road, the drivers behind find out late. ZnakOS tells them first: a signed warning, sent to nearby vehicles over the cellular network, with no roadside hardware. Shown here in simulation.
Request a briefing →It follows hazards that move.
Oversize loads, convoys and roadworks do not stay in one place. The warning travels with them, so each vehicle gets the right alert at the right distance. Shown in simulation.
Request a briefing →It speaks to riders, who cannot look at a screen.
A motorcyclist keeps their eyes on the road. ZnakOS gives a short spoken warning that needs no reply and no hands. Shown in simulation.
Request a briefing →Built, and entered in the RTA Dubai World Challenge 2027.
The platform is ready to deploy. We are waiting for the competition announcement. It is made for road authorities, operators, and connected and autonomous vehicles.
Request a briefing →Synapse is the intelligence layer of Nextwrks.
Every Nextwrks product runs on it: machine learning, perception and decision-making, built once and shared. It turns live data from roads, buildings and machines into decisions a person or a vehicle can act on. Behind it sits our research at the edge of neuroscience and computing.
ATHAR shows a city what is about to close.
Roadworks, utilities, events and construction are planned by separate teams, so no one sees all of it. Drivers and fleets find out at the barrier. ATHAR uses machine learning to bring every plan into one picture.
Request a briefing →One map of every planned work.
Planners and the public see what is closed, where and when. Vehicles can route around it before they reach the barrier.
Request a briefing →It catches clashes before crews deploy.
If two works would close the same junction on the same days, ATHAR flags it. Moving one job by a few days removes the problem. Launching within weeks.
Request a briefing →TENET.WORKS is a code-native design studio.
Brand, interface and engineering are built together, in code, so the design that is approved is the design that ships. TENET.WORKS is the studio behind the Nextwrks brands and the partners shown here.
Visit tenet.works →ZARC designs buildings as rules, then tests them.
It is the computational architecture division of Nextwrks: critical and secure buildings, smart buildings and digital twins. Every version of a design is tested against its brief before anyone draws a wall.
Security starts with the plan.
The layout decides who moves where and what can be seen. ZARC sets access, sightlines and exposure in the architecture of a secure facility, not in procedure added afterwards.
From a surface to a set of panels.
A free-form facade is cut into flat triangular panels, each with its own number, so it can be cut, built and fitted. One rule decides how every panel opens to the sun.
Nextlab builds robots that see before they move.
It is the robotics research division of Nextwrks. The first question is perception: how a machine detects what is around it and decides where it is safe to go. Everything is tested in simulation before it touches a floor.
A parts robot that plans its own route.
An autonomous robot for indoor sites. It plans a route, drives it, and carries parts from shelf to station with the arm on board. In design, and tested in simulation.
A robot that repairs a wall without being told where.
An autonomous construction robot that reads the plan, scans the wall, finds each marked repair point and fixes it. In design, and tested in simulation.
Tools for reading a place before it is built.
UrbanWrks is a set of connected tools for the architects, planners and researchers who study sites. Several tools, one platform. What each one measures stays private until launch.
Visit urbanwrks.com →NatureOS is open infrastructure for ecology.
Ecological data is scattered and closed. NatureOS is an open base for it: seven engines that can be tested, an open-source core, and AI agents that use the engines but never change them.
Visit natureos.site →The next systems are not public yet.
████ is a stealth research lab inside Nextwrks, working on ██████ ████████. Nothing about it is public until there is something to show. If you work on a related problem and want to build early, write to us.