Billions invested in AI—but how do we measure the return?

In Saudi Arabia, the question is no longer: Will we use AI?

The numbers suggest that adoption and investment are already underway. Government spending on telecommunications and information technology services reached around SAR 31.9 billion in 2025, while spending by government entities on AI and emerging technologies rose by 20%.

In the private sector, Saudi AI companies attracted around $9.1 billion across 70 deals in 2025, while the number of commercial registrations for AI-related activities exceeded 24,000 in 2026, growing by more than 31%.

These figures show the scale of the market taking shape, but they raise another question for me:

After all this investment, how do we know whether AI has created real value?

Investment is not the outcome

When a company announces that it has invested millions of riyals in AI, the figure itself may seem like a sign of progress.

But from a business perspective, the size of the investment does not tell us whether it was a good decision.

A company may spend on systems, infrastructure, data, training, and consulting, but real value starts when that spending makes a difference to its operations.

Has the cost of completing a particular task gone down?

Are employees more productive?

Has service delivery become faster?

Has the system helped increase revenue or reduce errors?

This is where we move from talking about the cost of technology to talking about its return.

AI has costs that are not always apparent at first

When considering AI, it is easy to focus only on the cost of buying or subscribing to a system.

But implementing it in practice may require technical infrastructure, data storage, cybersecurity, employee training, and continuous system monitoring, as well as the costs of updates and development.

That means an assessment should not stop at the initial cost.

A solution may seem cheaper to buy but become more expensive to operate. Another project may require a large upfront investment but save the company recurring costs for years.

So the financial question becomes: Will the benefits the project delivers over its lifetime justify the full cost of the investment?

And what about the risks?

The news is not just about investment; it also highlights Saudi Arabia’s ambitions in AI ethics and governance.

That matters, because a financial return does not make a project successful if it creates greater risks in the process.

An AI system may help reduce costs, while also creating risks related to data, privacy, or the quality of decisions.

That is why governance should not be a barrier to investment; it should be part of how investment is evaluated.

A sound decision does not ask only: How much will we save?

It also asks: What risks are we taking on in exchange for those savings?

From buying technology to putting it to use

What stood out to me most in the news was how Mohammed Al-Bishi, head of the Al-Eqtisadiah Forum, summed up this shift in the question itself: the question is no longer whether we use AI, but how we use and invest in it—and what this transformation costs and returns.

That is where I see the difference between adopting technology just to keep up with the market and using it as an economic decision.

A company does not need AI in every process just because its competitors use it.

In some processes, the technology may deliver major savings in time and cost, while in others, investing in it may not make economic sense at all.

So companies’ success in the next phase may depend less on who spends the most on AI and more on who knows where to invest.

What does an accountant look at?

When an accountant hears that the company plans to invest in an AI project, the only question will not be: How much will it cost?

There are other questions:

What are the upfront and ongoing costs? How much time or expense will the system save? Will it generate new revenue? When can the company recoup its investment? And what is the expected return compared with the alternatives?

Governance adds another dimension: Are there clear controls for using the system? And who is accountable for its decisions, outputs, and associated risks?

In my view, this is the most important shift.

AI may be a technology, but the decision to invest in it is ultimately a financial decision too.

The figures we see today—billions of riyals and dollars invested in the sector—tell us how ambitious the effort is, but they are not, on their own, a measure of success.

The real measure will be whether institutions can turn these investments into higher productivity, lower costs, new revenue, and better decisions—while managing the risks that come with them.

So perhaps the question is no longer:

How much are we investing in AI?

But rather:

How much value does every riyal we invest in it create?