When we use an AI model to create an image or write text, the process seems simple: we enter a prompt and receive the result within seconds.

But behind those seconds lies a vast economic and technological chain that begins with data centers, energy, and chips, passes through models and software, and ends with the skills and applications used by individuals and companies.

That is why building an AI sector does not mean merely owning a smart application or developing a single model. It requires building what can be called the “AI value chain.”

The recent partnerships between the Ministry of Communications and Information Technology and HUMAIN, along with the expansion of the ministry’s and HUMAIN’s partnership with Adobe, help illustrate how this chain takes shape in practice.

The First Layer: Computing Capacity

Before training any advanced model, we need something fundamental: computing.

Modern AI models require vast amounts of processing, which in turn requires specialized chips, servers, networks, and data centers capable of running them, along with the necessary energy, cooling, and digital infrastructure.

This is where the importance of AI data centers comes in.

Computing capacity has become akin to the infrastructure of traditional industries. Just as a factory needs electricity, roads, and ports, an AI economy needs data centers and computing capacity before it can develop applications and services on top of them.

Therefore, the cooperation between the Ministry of Communications and Information Technology and HUMAIN is not focused on a particular application, but on enabling the foundational infrastructure on which multiple applications can be built in the future.

Why Have Data Centers Become an Economic Issue?

Because the ability to run AI itself has become a valuable resource.

Companies that want to develop models or run AI services need to buy or rent computing capacity. If that capacity is available locally and at an appropriate scale, the market’s dependence on infrastructure located outside the economy can be reduced. The country could also become a base for hosting and operating services that serve other markets.

Here, the data center shifts from being a building containing servers to an economic asset capable of attracting companies, investments, and related digital services.

But computing alone does not create an AI economy.

The Second Layer: The Model

Above the infrastructure come the models that make use of this computing capacity.

This raises a different question: Is it enough to use a global model as it is, or do markets need models that better understand the local language, culture, and context?

The partnership between HUMAIN and Adobe offers an example through the development of an image-generation model designed to handle the Saudi cultural context and Arabic-language prompts.

This point is economically important because the value of AI comes not only from its technical capabilities, but also from how well its outputs suit the user who will benefit from them.

A globally powerful model may be less effective if it does not understand the language, cultural context, or needs of the market in which it operates.

This makes AI localization different from simply importing and using software.

The Third Layer: The User

Even the best infrastructure and the most powerful model create no economic value if no one uses them.

This is where another figure in the Adobe partnership becomes important: more than 27 million eligible citizens and residents aged 13 and above are expected to receive access to Adobe Firefly Standard and Adobe Express Premium features for 12 months.

The economic idea here goes beyond a free subscription.

Reducing the cost of access to the technology can lower what is known as the adoption barrier.

An individual who was not willing to pay to try AI tools may begin using them when they become available; a small company may introduce the technology into its operations; a student may acquire a new skill; and a creator may use it to increase productivity.

The broader the user base, the more likely the technology is to evolve from a limited experiment into a tool embedded in everyday economic activity.

What Does the $4 Billion Figure Mean?

Adobe announced that the value of its commitment under the partnership exceeds $4 billion.

But it is important to understand the figure correctly.

This figure does not necessarily mean that $4 billion in cash will be injected into a project or that assets of the same value will be built. Rather, it represents the value the company attributes to its commitment under the program, which includes making its services and tools available to millions of users.

This is an important point when interpreting economic announcements in general:

The value of an agreement, the value of services provided, and the amount of capital investment are not necessarily the same thing.

Distinguishing between them helps clarify the actual economic impact rather than focusing solely on the size of the announced figure.

The Fourth Layer: Human Capital

There is another element that cannot be bought simply by building a data center: knowledge.

That is why cooperation plans include transferring expertise, building national capabilities, and providing training.

The reason is simple: the more an economy has engineers, developers, and researchers capable of building, operating, and improving technologies, the greater the share of added value tied to the local economy.

There is a significant difference between an economy that consumes AI services and one that has the talent to develop, operate, and export those services.

This is where the importance of training becomes clear—not as a social dimension separate from investment, but as part of the capital the sector needs in order to grow.

The Ecosystem Matters More Than Any Single Technology

The AI economy can be viewed as an interconnected chain:

Data centers provide computing,
computing runs models,
models enable applications,
applications need users,
and users and companies need the skills to benefit from them.

If one of these links is missing, the value that can be extracted from the others is reduced.

A data center without applications remains an underutilized capacity; an advanced model without users generates no broad economic impact; and making tools available without the necessary skills may not lead to a real increase in productivity.

That is why the most important question when evaluating AI investments is not simply: How many billions were announced?

Rather:

How much computing capacity will be built?
How much of it will actually be used?
How many models or products will be developed on top of it?
How many companies will adopt these technologies?
How many skills, jobs, and sources of added value will emerge as a result?

This is where the bigger picture emerges.

The AI race is not only about who has the most advanced application, but about who can build the chain that transforms computing capacity into knowledge, knowledge into applications, and applications into productivity and economic value.