By Lara Tseng
Most of the conversation about artificial intelligence still happens on screens. We argue about whose model is biggest, who has the most chips and what the latest chatbot can write. Those questions matter, but they miss the part of AI I think will shape the next decade: what happens when it has to operate in the physical world.
Put AI into a port, a factory, a road network or a power grid and the job changes. A plausible answer is no longer enough. The system has to read what is happening around it, react in fractions of a second and get it right, over and over, in conditions nobody fully controls.
That is why I believe the Middle East, and the UAE in particular, could become one of the world’s most important proving grounds for AI. The region is doing more than talking about AI. It is building the infrastructure where AI can be deployed, pushed to its limits and improved at scale.
Anyone who has put technology into the field knows the gap between a demo and daily operation. You can stage a demo, but you can’t stage real life. Lenses collect dust, temperatures swing, networks clog, machines drift as they age and people behave in ways no dataset predicted. Those conditions separate AI that is useful from AI that only looks good in a slide deck.
The UAE has spent years putting technology into exactly these environments: transport, logistics, energy and public services. These sectors decide how quickly people get to work, how goods move and whether the lights stay on.
Mobility is the clearest example. Dubai wants a quarter of all trips in the city to be autonomous by 2030. Getting there takes far more than putting driverless cars on the road. Each vehicle has to make sense of pedestrians, signals, lane markings and other drivers, and decide what to do almost instantly.
That decision can’t always travel to a data center hundreds of kilometers away and wait for a reply. Some of the intelligence has to sit where the action is. This is the case for edge AI: a camera that analyses what it sees on the spot, a machine that flags an abnormal vibration before it breaks down, a vehicle that reacts to the road without waiting for instructions. The cloud still matters for training models and seeing the bigger picture. The systems that succeed will use both, and the real design question is how to divide the work between them.
Logistics may be where this plays out first. The region sits on some of the world’s busiest trade routes, and its ports, free zones and warehouses already generate enormous amounts of data. Trucks are always moving, containers are tracked from quay to warehouse, and equipment is expected to run around the clock.
AI already helps with parts of this. It can predict when a crane motor is about to fail, check cargo with computer vision and reroute vehicles around congestion. The bigger prize is connecting those pieces. Then a port has a live picture of its own operations, a warehouse adjusts as demand shifts, and a transport network responds to what is happening now instead of following yesterday’s plan.
That points to a different way of building. For most of the last century we built roads, factories and ports first and added technology later. Much of the Middle East is still building new districts, rail lines, economic zones and even entire cities. That gives it something mature markets rarely have: the chance to design infrastructure with intelligence built in from the start instead of bolted on afterwards.
Energy deserves more attention in this debate than it gets. AI runs on computing power, and computing power runs on electricity. As data centers multiply, energy will increasingly decide where and how AI can grow.
The Middle East brings decades of energy expertise to a moment when it is also investing heavily in digital infrastructure. But bigger data centers can’t be the whole answer. AI also has to become more efficient. Not every sensor reading needs to cross a continent before someone acts on it. Processing data closer to where it is generated cuts unnecessary traffic, saves power and speeds up response times.
In some sectors, response time is everything. A delay in a factory can damage equipment or halt a production line. In a utility network it can mean an outage. In an autonomous vehicle, the stakes are higher still.
Other regions have their own strengths. The United States leads in AI research and investment, East Asia has world-class manufacturing, and Europe has deep engineering talent and some of the most developed thinking on regulation. The Middle East offers something different: large infrastructure projects, governments directly involved in delivering them, and a willingness to move quickly from plan to implementation. There is also a real appetite for new technology, from governments modernizing public services, companies looking to automate and cities competing for investment and talent.
Not every experiment will work. Some will fail, some will need rethinking, and some will raise hard questions about data, security and public trust. That is how useful technology gets made.
Speed must not come at the expense of responsibility. Once AI starts steering physical infrastructure, a mistake means more than a bad paragraph. It can mean a stalled production line or a vehicle that brakes too late. Security, privacy and governance have to be designed in from the start. But those safeguards can’t be perfected in theory. They have to be tested, carefully, in the places where these systems will actually run.
The Middle East doesn’t need to invent every new AI model to matter in this story. Its contribution could be just as valuable: showing the world how AI performs once it leaves the lab.
Lara Tseng is CEO of Vecow for the Middle East, North Africa and Türkiye, where she focuses on edge AI and the practical deployment of AI across industry and infrastructure.
