In 2025, the Arab world entered a different phase in its engagement with artificial intelligence. The discussion was no longer about “catching up” or “encouraging innovation,” but about massive figures, mostly in the tens of billions of dollars, put on the table to build a robust digital infrastructure comparable, in its logic, to oil and gas investments in the last century.

The question that imposes itself is: What will these billions actually mean in 2026? And what will we discover when artificial intelligence moves from the announcement phase to the operational phase?

The figures announced in 2025 show a clear trend: The bulk of the funds was not directed toward applications or startup but toward the “solid foundation” of the AI economy, namely computing power, data centers, and advanced cloud infrastructure. In this context, the progress of the United Arab Emirates is clear evidence of owning the infrastructure assets themselves, rather than merely hosting applications.

Through the MGX fund, an Abu Dhabi–backed investment vehicle focused on advanced technologies, Abu Dhabi moved toward investing in globally scaled data centers. Talk of a deal worth nearly $40 billion linked to Aligned Data Centers reflects a strategic orientation based on “physical” control that is, buildings, energy, and cooling as a prerequisite for influence in the global AI economy. Reports that the same fund is seeking to raise its investment volume to as much as $25 billion further confirm that competition is no longer merely regional, but has become part of an international race over AI infrastructure.

Saudi Arabia, in turn, has chosen a parallel path. During the LEAP 2025 conference, new commitments and projects totaling around $14.9 billion were announced in artificial intelligence and digital infrastructure, covering advanced computing, semiconductors, robotics, and data centers. These figures reflect a clear strategic decision: build capability before building the market.

This decision became even more explicit in May 2025 with the launch of the Humain platform under the umbrella of the Public Investment Fund (Saudi Arabia’s sovereign wealth fund), as a national arm for AI development based on the “full stack” approach, from data centers and a sovereign cloud to language models, including multimodal Arabic models, and onward to solutions tailored for government and the private sector.

As for Qatar, it entered the race at a relatively later stage, but with a large figure. In December 2025, a partnership worth nearly $20 billion was announced with Brookfield, a global asset management firm, to build an advanced computing platform and data centers.

What does all this mean in 2026? First, there will be a shift from an “announcement economy” to an “operations economy.” Success will no longer be measured by the number of conferences or the size of announced commitments but by the number of megawatts actually brought online, occupancy rates, and the length of leasing contracts. Every billion dollars announced in 2025 will be tested in 2026 by a simple question: What proportion has been converted into functioning computing capacity?

Second, the first generation of Arab AI services will emerge at scale, particularly in the government, financial, and telecommunications sectors. However, this emergence will remain limited in impact unless one of the biggest unresolved unknowns of 2025 is addressed: the availability of advanced chips and the geopolitical constraints tied to them. Infrastructure without sufficient processors could turn into a costly burden.

Thirdly, the question of local demand will arise. Will Arab governments and institutions become permanent customers of AI services, or will reliance remain on external demand to absorb surplus capacity? The answer to this question will determine how many of the billions of 2025 will lead to genuine economic transformation, and how many will result in a model that exports computing capacity.

Ultimately, 2026 will be the year when Arab AI billions crystallize into a tangible reality. Not because it is a year of launches, but because it is a year of measurement: measurement of operations, usage, and added value. Only then will we know how many of the figures announced in 2025 laid the foundations for a new economy, and how many will bear fruit at a later stage.