Saudi Arabia’s push into artificial intelligence and data centers is adding another enormous financing requirement to an economy already trying to fund the costly ambitions of Vision 2030.
According to Semafor, HUMAIN, the Public Investment Fund-owned AI company, is targeting more than 6 gigawatts of data center capacity over the next decade. But financing just half of those plans could require as much as $32 billion in debt. That is roughly 120 billion Saudi riyals.
The more revealing finding, however, comes from consulting firm Alvarez & Marsal: Saudi banks would not be able to provide that amount of financing on their own.
This turns Saudi Arabia’s AI story into something more than a race for technology. It raises a much more basic question: who will finance the kingdom’s ambitions when the projects become larger than the domestic banking system’s ability to fund them?
6 Gigawatts Comes With a Huge Price Tag
Data centers make sense as part of Saudi Arabia’s diversification strategy.
Global demand for computing power is rising rapidly with the growth of artificial intelligence. Saudi Arabia also has several advantages: access to energy, available land, significant state capital, and a government determined to establish the kingdom as a major regional technology hub. But data centers are extremely expensive to build.
Long before a facility can sell computing capacity and generate revenue, billions must be spent on land, construction, power connections, cooling systems, networking infrastructure, servers, and other specialized equipment. That is why the $32 billion estimate matters.
It would finance only around half of HUMAIN’s targeted capacity. If even half of one company’s expansion plans requires tens of billions of dollars in borrowing, the question is no longer simply whether Saudi Arabia has the ambition to become an AI power. It is whether it can finance that ambition without adding another major layer of debt to Vision 2030.
Saudi Banks Cannot Carry the Bill Alone
Perhaps the most important point is that Saudi Arabia’s domestic banks are not expected to be able to finance the required debt by themselves.
That means HUMAIN would likely need a much broader pool of lenders and investors, potentially including international banks, private credit funds, global debt markets, and other foreign financial institutions.
And here, once again, Vision 2030 runs into a familiar contradiction.
Saudi Arabia launches new industries to diversify its economy and create new sources of income. But building those industries requires increasingly large amounts of capital, borrowing, and outside financing before they generate meaningful returns.
Diversification is supposed eventually to create new sources of wealth.
Getting there, however, is becoming extraordinarily expensive.
HUMAIN Is Not Competing for Capital Alone
If HUMAIN were the only Saudi project seeking tens of billions of dollars, the financing challenge might be relatively contained. It is not.
Saudi Arabia is simultaneously trying to finance NEOM and its various developments, Qiddiya, Diriyah, tourism projects, airports, transport infrastructure, energy projects, mining expansion, preparations for the 2034 FIFA World Cup, and numerous companies backed by the Public Investment Fund.
The government itself also borrows. So does the PIF. So do many state-backed companies. All of them ultimately compete for capital. That is where the real pressure begins.
Saudi banks do not have unlimited balance sheets, and international lenders do not provide billions of dollars without demanding returns, guarantees, favorable terms, or confidence that the projects they finance can generate enough cash to repay them.
The larger Saudi Arabia’s investment pipeline becomes, the more important the cost and availability of capital becomes.
AI’s Future Is Being Built With Very Traditional Debt
Kurt Davis Jr., who authored the Alvarez & Marsal report cited by Semafor, described digital infrastructure as one of the biggest new sources of project debt in the firm’s portfolio.
That observation captures an important part of the AI boom that often disappears behind futuristic announcements.
Artificial intelligence may be marketed as the industry of the future, but the infrastructure behind it depends on something much more traditional: debt.
Data centers are not built with speeches about technological leadership.
They require billions of dollars from someone willing to lend the money. Those loans carry interest, repayment schedules, and financial risks. The facilities must eventually generate enough revenue not only to operate, but also to service the debt used to build them. The technology may be new. The financial mathematics is not.
Another Giant Bill for the Public Investment Fund
HUMAIN is owned by the Public Investment Fund, placing its expansion inside an already enormous portfolio of Saudi bets.
The PIF is investing across tourism, real estate, sports, gaming, electric vehicles, aviation, mining, technology, and other sectors while simultaneously financing some of the kingdom’s largest domestic developments.
Artificial intelligence and data centers now add another capital-intensive industry to that list.
Diversification itself is not the problem. Investing across different sectors is precisely what a sovereign wealth fund can be expected to do.
The problem begins when many of those sectors require enormous amounts of capital at the same time. At some point, priorities have to be set. Capital is not unlimited, even for Saudi Arabia.
When Saudi Banks Are Not Enough, Foreign Lenders Enter
There are obvious ways to fill the financing gap. HUMAIN could seek syndicated international loans, issue bonds, attract private credit, bring in foreign partners, or combine several financing structures. But none of that money comes free. Foreign investors expect returns. Banks charge interest. Bondholders expect repayment. Private credit can be particularly expensive.
And the greater the perceived risk, the more expensive that capital becomes. Saudi Arabia could therefore find itself in an unusual position: borrowing internationally to build the infrastructure of an economy that is supposed to become more diversified and economically independent in the future.
There is nothing inherently wrong with financing productive infrastructure through debt. Companies and governments around the world do it constantly.
But the model only works if the future economic returns justify the cost of the capital used to build it. That remains the critical test.
The Number That Matters Is Not 6 Gigawatts
Saudi Arabia can announce 1 gigawatt, 6 gigawatts, or even more. But installed capacity will not determine whether HUMAIN succeeds. Utilization will.
Who will actually use these data centers?
How much will customers pay?
How long will their contracts last?
Can Saudi Arabia attract global technology companies and AI developers to use this capacity consistently?
And will revenue be high enough to cover construction, electricity, cooling, hardware, maintenance, and billions of dollars in debt?
If the answer is yes, borrowing could help build a valuable and profitable new industry.
If infrastructure is built much faster than real demand develops, however, Saudi Arabia could be left with extremely expensive assets that continue consuming capital instead of generating it.
That is the risk hidden behind the headline gigawatt figures.
AI Is Not a Guaranteed Return
There is another problem: the global AI infrastructure boom itself is moving extraordinarily quickly.
Demand for computing capacity is currently enormous, but so is investment.
Chip technology is changing rapidly. AI models are becoming more efficient. Cooling and energy technologies are evolving. Technology companies are constantly reassessing how much computing infrastructure they will need and where they want it located.
Saudi Arabia is therefore not entering a mature and predictable industry.
It is entering one of the fastest-moving and most competitive investment races in the world.
Spending tens of billions of dollars based on assumptions about future AI demand is a major bet. If those assumptions prove correct, the returns could be significant.
If they do not, the debt will remain even if the demand does not.
Vision 2030 Is Running Into a Capital Problem
The larger issue goes far beyond HUMAIN.
Almost every Saudi mega-project eventually reaches the same question:
Who pays for it?
During the early years of Vision 2030, the answer was often the government or the Public Investment Fund.
But as the list of mega-projects has expanded, capital can no longer be treated as if it has no limit.
If financing only half of one company’s data center plans could require $32 billion in debt—and domestic banks cannot provide that financing alone—it offers a glimpse of the pressure created by an economic transformation involving dozens of huge projects at the same time.
Saudi Arabia is no longer simply deciding what it wants to build.
It increasingly has to decide what it can afford to build, what should be financed first, and how much debt the transformation can reasonably carry.
The Digital Future Comes With a Very Real Bill
None of this means HUMAIN’s data center strategy is destined to fail.
Saudi Arabia has genuine advantages that could help it build a major AI infrastructure industry, and global demand for computing capacity represents a real economic opportunity.
But the financing estimates reveal the part of the story that futuristic announcements often leave out.
Six gigawatts is not simply a technology target. It is an enormous financial commitment.
And $32 billion is not just a figure in a presentation. It represents money that must be borrowed, interest that must be paid, and infrastructure that must eventually generate enough cash to justify the investment.
More importantly, that estimate covers only around half of HUMAIN’s planned capacity.
The real test for HUMAIN will therefore not be whether Saudi Arabia can announce enormous data centers or buy the latest AI chips.
It will be whether those facilities can eventually produce more money than they consume.
Saudi Arabia can buy servers. It can build data centers. And, with the right terms, it can probably attract international lenders willing to finance them.
What it cannot buy in advance is sustainable economic demand.
That leaves Vision 2030 facing an increasingly familiar question: if even Saudi Arabia’s own banks cannot finance the scale of its ambitions, how many more mega-projects can the kingdom continue adding before financing the transformation itself becomes one of the biggest limits on it?
