The Line is dead. Not officially — Saudi Arabia would never say that — but NEOM has halted all construction on The Line until after 2030, and if you understand what “after 2030” means in Saudi megaproject speak, you know the translation is “probably never.” The 170-kilometer mirrored skyscraper that was supposed to house 9 million people in a car-free, carbon-neutral corridor through the desert? It’s been downgraded to a population target of 100,000. That’s not a city. That’s a suburb of Jeddah with better marketing.
But here’s where it gets interesting: NEOM isn’t dead. The money is just going somewhere else. Specifically, it’s going into AI data centers — a $5 billion partnership with DataVolt to build a massive AI compute campus in NEOM’s Oxagon industrial district. Saudi Arabia looked at its $1 trillion vanity project, looked at the AI boom eating the world’s electricity supply, and made the coldest calculation in the history of sovereign wealth funds: server racks generate better returns than mirror walls.
The Line Was Always a $1 Trillion Bet Against Physics
Let’s be honest about what The Line actually was: the most expensive piece of concept art ever commissioned. A pair of parallel mirrored skyscrapers, 500 meters tall and 200 meters wide, stretching 170 kilometers through the Saudi desert. No cars. No roads. No carbon emissions. A high-speed train running the entire length. Vertical farms feeding millions. Total projected cost: over $1 trillion.
The problem wasn’t just money — it was thermodynamics. Building a 170-kilometer structure in a desert where temperatures hit 50°C, with no proven construction methodology for anything remotely like it, was always going to collide with reality. And reality showed up faster than anyone expected. Widening budget deficits, lower-than-expected foreign investment, the economic shock from regional instability — all of it forced the Public Investment Fund to start asking a question it had been avoiding: what if we spent this money on something that actually works?
The AI Data Center Pivot Is the Quiet Part Said Loud
Here’s what nobody is saying about this pivot: Saudi Arabia just admitted that the future isn’t futuristic cities — it’s boring infrastructure. The kingdom looked at where the actual demand is in 2026 and saw the same thing everyone else sees: every major AI company on the planet is desperately hunting for places to put GPUs. Microsoft, Google, Amazon, Oracle — they’re all building data centers as fast as they can, and they’re running out of places with cheap electricity, cool climates, and governments willing to look the other way on power consumption.
Saudi Arabia has two of those three. The desert isn’t cool, but it has effectively unlimited cheap energy and a government that will bulldoze regulatory obstacles faster than any democracy on Earth. The $5 billion DataVolt partnership in Oxagon is the opening move, but the play is much bigger: position NEOM as a hyperscale AI compute hub that competes with Northern Virginia, Singapore, and the Nordics for the next generation of AI infrastructure spending.
And the numbers make the argument by themselves. The global data center market is projected to exceed $500 billion by 2030. AI workloads are doubling every six to nine months. Every major cloud provider has announced plans to spend between $50 billion and $100 billion on AI infrastructure this year alone. That’s real money chasing real capacity. The Line was a render. Data centers are revenue.
What This Actually Means for the AI Industry
If Saudi Arabia executes on this — and execution is always the asterisk with NEOM — it could materially change the geography of AI compute. Right now, AI infrastructure is concentrated in the United States, with secondary hubs in Western Europe and Southeast Asia. The Middle East is a gap on the map. A massive, sovereign-backed compute campus with cheap power and favorable tax treatment could pull workloads from companies that don’t want to be entirely dependent on American or Chinese infrastructure.
Think about who benefits from this: AI companies with sovereignty-sensitive customers. European companies that don’t want their data on American soil. Asian companies that need a neutral third option. Government clients across the Middle East and Africa. There’s a genuine market here — and Saudi Arabia is the only Gulf state with both the capital and the physical space to build at the scale the AI industry requires.
The DataVolt deal signals the model: partner with operators who bring the technical expertise, provide the land and power and regulatory fast-track, and take a cut of the compute economics. It’s the same playbook Abu Dhabi used to build a global aviation hub with Emirates and Etihad — except instead of airports, it’s GPU clusters.
The Uncomfortable Truth About Vision Projects
The Line was always more about Mohammed bin Salman’s brand than about urban planning. It was a recruiting tool for foreign talent, a negotiating chip for diplomatic meetings, and a signal to the world that Saudi Arabia was serious about a post-oil future. And on those metrics, it worked — for a while. But vision projects have a half-life, and The Line’s expired the moment the first budget cut landed.
The pivot to AI infrastructure is the opposite of a vision project. It’s unglamorous, utilitarian, and exactly what the market is paying for. Nobody is going to make a documentary about a data center campus in the Saudi desert. But the revenue from hosting Nvidia H200 clusters for Anthropic or Google is infinitely more real than the revenue from a mirrored city that was never going to exist in the first place.
That’s the lesson MBS is learning — the same one every founder eventually learns. The market doesn’t care about your vision. It cares about your execution. And right now, the market is begging for more GPU racks, not more concept renders.
The Verdict
Saudi Arabia just traded the most ambitious architectural project in human history for a bunch of air-conditioned warehouses full of servers. And it’s the right call. The Line was a monument to ambition. The data center pivot is a monument to arithmetic. In 2026, arithmetic wins every time.
The real question isn’t whether this pivot makes financial sense — it obviously does. The real question is whether Saudi Arabia can build data centers at hyperscale as well as it builds marketing campaigns. Because so far, NEOM’s greatest export has been CGI renders, not kilowatt-hours. The $5 billion DataVolt deal is a start. But between a press release and a functioning AI compute campus, there are about a million engineering problems that don’t respond to sovereign wealth.
Still: would you rather be the country building a mirror in the desert, or the country selling compute to every AI company that can’t find a power outlet? MBS just chose door number two. And for the first time in NEOM’s history, the math actually works.