When we talk about the artificial intelligence race, attention usually turns to the most powerful models and fastest chips. But as the technology’s use expands, a less familiar factor with stronger economic implications is emerging: How much does it cost to run artificial intelligence, and where can the computing capacity it requires be provided?

This perspective helps explain why energy and data centers have become part of the global competition over artificial intelligence, and why Saudi Arabia could be a player at this stage even without being a global center for chip manufacturing.

The Chip Alone Is Not Enough

Artificial intelligence chips do not operate in a vacuum. Running them at scale requires data centers, reliable power grids, cooling and communications systems, as well as massive investments in infrastructure.

These elements are becoming increasingly important as demand for computing grows. The International Energy Agency expects global electricity consumption by data centers to rise from approximately 485 terawatt-hours in 2025 to around 950 terawatt-hours in 2030, while electricity consumption by data centers dedicated to artificial intelligence is growing at a faster pace.

This is changing the nature of competition. Owning advanced chips remains important, but the ability to operate them efficiently, reliably, and at scale has become part of the economic equation itself.

From Chip Speed to Operating Cost

After an artificial intelligence model is trained, a phase called inference begins. Simply put, this means running the model each time a user requests an answer, image, or task.

With millions of users and requests, each small operation becomes a recurring operating cost. That is why chip companies have begun looking at metrics such as Cost per Token, rather than chip speed alone.

South Korean company Rebellions, which specializes in artificial intelligence chips, focuses specifically on inference chips. It says competition will increasingly center on delivering good performance at a lower operating cost, directly linking chip efficiency to the energy and infrastructure that power them.

Economically, the idea is simple: as artificial intelligence shifts from a technology experiment to a service that companies use continuously, daily operating costs become more important than the cost of building the model alone.

Where Does Saudi Arabia Come In?

This is where Saudi infrastructure takes on significance beyond simply building data centers.

The Ministry of Communications and Information Technology views data centers as a fundamental asset in developing the Kingdom as a digital hub, with long-term plans to expand local computing capacity. Ministry documents also indicate that the availability and cost of energy are factors supporting the Kingdom’s attractiveness for data center investments.

That is why Rebellions’ decision to expand in Riyadh is directly linked to the economics of operations. The company believes that energy availability, investment in artificial intelligence infrastructure, and the move toward Sovereign AI are factors supporting its presence in the Kingdom.

Sovereign AI does not mean that a country must produce all the technology itself. Rather, it means having greater control over where data is stored, where models are operated, and the infrastructure they rely on, particularly in sensitive sectors.

The Opportunity Is Greater Than Importing Chips

Saudi Arabia’s economic opportunity does not necessarily lie in directly competing with South Korea or Taiwan to manufacture the most advanced chips. Another part of the value chain can be built locally: operating computing infrastructure.

Having chips inside Saudi data centers could create activity around operations, cloud services, cybersecurity, networks, maintenance, and artificial intelligence applications.

But building data centers alone does not guarantee the realization of this value. Success depends on genuine demand for computing, reliable electricity capacity, and local skills and companies capable of building services and products on top of this infrastructure.

This is where the difference emerges between owning infrastructure and building an economy around it.

The Next Race May Be About the Cost of Computing

Artificial intelligence is reshaping the technology value chain. The question is no longer simply: Who makes the best chip? It is also: Who can operate it at the lowest cost, with the greatest reliability, and closest to users and data?

For Saudi Arabia, abundant energy and the expansion of data centers could provide an important entry point into this chain. But the real test will be the Kingdom’s ability to transform this infrastructure from facilities that host servers into an ecosystem that generates local companies, skills, services, and economic value around it.