Every country is starting to put AI in some form as part of their national strategy. It’s also creating an uneven playing field on tracking its adoption worldwide. The Microsoft AI Economy Institute has just published the “AI Diffusion Report” (source) that looks to provide a deeper look. It’s a comprehensive look at where Artificial Intelligence is being used, developed, and built across the globe. Artificial Intelligence is the fastest-spreading technology in human history, having reached over 1.2 billion people in under three years.
To clear out any confusion to the term “AI Diffusion”, it covers the adoption of AI and adaptation to its use. What stands out from this report will be a bright spot for the Middle East region.
UAE is a frontrunner in AI adoption
A major highlight from the report is a big boon for one country in the Middle East region. The United Arab Emirates (UAE) as a global leader in the adoption of AI technologies. An AI Diffusion rate of 59.4% for their working-age adults using AI places UAE with highest rate of AI diffusion in the world.
This places the UAE ahead of other leading nations such as Singapore (58.6%), Norway (45.3%), and Ireland (41.7%). This can be attributed to sustained, long-term investments in developing digital connectivity and skills among the population.
The country is not yet at the technological frontier with regards to building its own models. Yet they have built their own Arabic LLM. It also showcases the disparity between the North and South with regards to AI adoption.
Wider MENA region
One thing that comes out from the report is the regional overview of AI diffusion. With the UAE already commanding a big lead, the next closest country in the region is Qatar with an AI Diffusion rate of 35.7%. What about the rest of the region?
Looking at the above scatter plot shows more countries from the region. AI adoption correlates strongly with GDP. Countries in the Global South, especially those with GDP per capita below $20,000, generally show slower adoption. Having perused the report’s tabulated results, I thought it best to illustrate them into a nice simple chart. Each chart is divided between the Middle East and North Africa regions.
Infrastructure capacity challenges
Running all the computationally-intensive AI workloads does require data centers equipped with high performant chips. They also have to be distributed worldwide to help share the workload and cut down on latency. A major observation is that most of the data centers are concentrated in Europe and USA/Canada. That’s courtesy of a steady and reliable source of electricity powering them.
Data centers are not a problem in the Middle East, as many of the big tech have set up regions. This includes the likes of Google Cloud, AWS, Microsoft Azure, and many more. The major part is that they’re divided between four main nations: UAE, Saudi Arabia, Qatar, and Bahrain (only AWS). Yet, it’s only the first three that are significant in this report.
The entire Middle East region is estimated to have an installed data center capacity of 1.1GW (gigawatts). That will go up with the expected launch of Stargate UAE sometime in 2026 for the first phase. The abundance of solar energy during the summer period is a gold mine to address the energy demands for powering them. As for North Africa, the closest comes from the South with a 1.5GW capacity.
Final Remarks
There are some parting words to wrap up this overview. Strong access to technology, education, and policy coordination can drive rapid adoption even without frontier-level model development or data centers.
If you’re still interested in reading the full report, you can view it here.
Image Credit: Brad Smith (Header Image) and Microsoft (other images)
