The Megaproject That Shrank
NEOM was announced in 2017 as Saudi Arabia’s most audacious Vision 2030 project: a $500 billion futuristic city on the Red Sea coast, designed from scratch as a technology-first environment that would showcase Saudi Arabia’s ability to reimagine urban civilization. The Line — a 170-kilometer linear city, 500 meters tall, 200 meters wide, housing up to 9 million people in a carbon-neutral built environment — was the project’s signature concept: a repudiation of traditional city planning that generated global attention and considerable skepticism in equal measure.
By 2024, the skepticism had been substantially vindicated. The Line had been scaled back from 170 kilometers to a 2.4-kilometer initial segment. Workforce reductions at NEOM’s development company had reached the thousands. Timeline commitments made in public presentations had slipped by years. The $500 billion headline figure had never corresponded to committed construction contracts, and the actual capital deployed toward construction remained a fraction of the announced budget.
For investors and vendors tracking NEOM as an AI infrastructure opportunity, the honest analysis requires both acknowledging the scaled-back reality and identifying what remains genuinely valuable. NEOM as originally envisioned is not being built on the announced timeline. NEOM as a technology development zone with specific, high-value AI infrastructure components — most importantly the DataVolt facility at Oxagon and the Tonomus technology mandate — is still a real program with real capital behind it.
The DataVolt Facility: Physical Anchor
The most credible AI infrastructure component at NEOM is the DataVolt data center project at Oxagon — NEOM’s industrial and logistics hub on the Red Sea coast. DataVolt announced a 1.5-gigawatt data center development at Oxagon, at an estimated cost of $5 billion, with a net-zero operational commitment backed by renewable energy from the region’s exceptional solar resources.
1.5 GW is an extraordinary scale. For reference, the largest hyperscaler data center campuses in operation globally are in the 500 MW to 1 GW range; a 1.5 GW facility would be among the largest AI-capable data center deployments in the world if fully built. The net-zero commitment is achievable in the NEOM context: the Tabuk region, where NEOM is located, has among the highest solar irradiance in the world, and NEOM’s early infrastructure development has prioritized renewable energy generation.
The DataVolt deployment represents NEOM’s most concrete AI infrastructure commitment. Unlike The Line’s architectural ambitions — which require solving problems of vertical transportation, air circulation, and social organization that have no tested precedents — the DataVolt facility is a conventional (if large) data center construction project with established engineering methodologies. The 1.5 GW target may take a decade to reach, but the initial phases of construction are proceeding, and the first operational capacity is expected to come online within the near term.
The customers for DataVolt’s capacity are the hyperscalers and AI companies that have signed into the broader Saudi AI buildout: the same Google Cloud, AWS, and Humain partners that are deploying AI infrastructure across multiple Saudi sites. Oxagon’s combination of available land, renewable power, and Red Sea cooling potential makes it a competitive location for large-scale AI compute relative to inland Saudi sites.
Tonomus: The Technology Mandate
Tonomus is NEOM’s technology subsidiary, responsible for all digital infrastructure, AI systems, and smart city technology across the NEOM zone. Tonomus was created to be the operational technology arm that transforms NEOM from a construction project into a functioning intelligent environment.
Tonomus’s mandate is comprehensive: designing and deploying the sensor networks, data platforms, AI inference systems, and connectivity infrastructure that will make NEOM’s smart city claims operational. This includes AI systems for real-time city management — traffic optimization across The Line’s internal transportation network, energy management for the building systems, water system monitoring, and the security and access control systems that a purpose-built city requires.
The OpenAI partnership for NEOM generative AI services runs through Tonomus. The practical scope of the OpenAI relationship at NEOM includes AI-powered resident services — navigation, information access, government service interfaces — that require generative AI capabilities. Tonomus is the entity that translates NEOM’s high-level technology vision into deployed systems.
Tonomus has faced the same scaling challenges as NEOM’s construction program. The original staffing plan called for a technology workforce of several thousand by 2024; actual hiring has been more modest, and the scope of what Tonomus is currently building reflects the reduced physical footprint of NEOM’s construction rather than the original vision. The 2024 workforce reductions at NEOM affected Tonomus as well as the construction and project management functions.
The surviving Tonomus mandate is still substantial. Even a NEOM at 10% of its original scale represents a significant technology deployment: a residential and commercial community of tens of thousands of people, with industrial operations at Oxagon, connected by new transportation infrastructure, operating in an extreme climate environment that places specific demands on energy and water systems. The AI systems required to manage that environment are real and valuable regardless of whether The Line ever reaches its full 170-kilometer extent.
The Greenfield Advantage: Why NEOM Matters for AI
The strategic insight behind NEOM’s AI relevance is not its scale but its architecture. NEOM is a greenfield environment — it was empty desert before construction began, and every system is being designed from scratch, without the constraint of legacy infrastructure.
Most smart city AI projects fail not because the AI is insufficient but because the city already exists. Retrofitting sensors into buildings designed in the 1960s, integrating AI traffic management with street layouts designed before autonomous vehicles existed, adding AI-powered energy management to grid infrastructure with decades of incremental investment — all of this is technically possible but operationally complex and expensive. NEOM has none of these constraints.
At NEOM, the sensor network is designed into the building systems from the beginning. The data flows are architected before the first resident arrives. The AI inference infrastructure is provisioned before the buildings are occupied. This allows a level of data richness and system integration that retroactive smart city programs cannot achieve. If NEOM is eventually populated at significant scale — even a fraction of the original 9 million target — the AI systems deployed there will have access to more comprehensive operational data, more real-time than any comparable built environment in the world.
This greenfield advantage is why the AI technology companies and research institutions that have committed to NEOM partnerships have done so despite the scale reductions and timeline delays. The opportunity is not scale — it is architecture. NEOM, even at its current reduced scope, represents the world’s most comprehensive real-world AI deployment testbed. The data generated by operating a purpose-built city with comprehensive sensing, AI energy management, and AI transportation systems will be scientifically and commercially valuable regardless of how many people live there.
The Brutal Contrast: Vision vs. Reality
The gap between NEOM’s announced ambitions and its current execution state deserves direct analysis for the investors and vendors making decisions about NEOM-related commitments.
The $500 billion figure was always a decade-long commitment across multiple components, not a near-term construction budget. The actual capital committed to construction contracts by 2025 is estimated at $20-40 billion — real money, but approximately 5-8% of the headline figure. The Line’s 2.4-kilometer initial segment will house thousands of residents, not millions. The architectural models showing gleaming glass towers extending to the horizon describe a decades-long vision, not a current project.
The 2024 workforce reductions were significant: NEOM’s headcount was cut by several thousand positions across construction management, design, and operations. Some of these cuts reflected the scaling back of The Line; others reflected cost discipline imposed by PIF’s broader portfolio management, which has been under pressure to show returns across its $700 billion in assets under management.
For AI vendors and technology partners who committed to NEOM projects in 2022 and 2023 — when The Line’s full vision was still being publicly promoted — the reduced scope creates commercial risk. Projects scoped against a 170-kilometer linear city have a different financial profile when scaled to 2.4 kilometers. Contracts that assumed NEOM would need AI systems for 9 million residents need renegotiation for a community of tens of thousands.
The honest vendor assessment is that NEOM-specific AI opportunities are real but must be re-scoped against current construction commitments rather than announced visions. The DataVolt facility and the Tonomus technology mandate survive the scale reduction because they are anchored to physical infrastructure — data centers and operational technology systems — rather than to the residential population that The Line was supposed to attract.
NEOM in Saudi Arabia’s AI Strategy: Showcase vs. Industrial AI
NEOM has always occupied a somewhat separate space within Saudi Arabia’s AI strategy from the industrial and government AI programs that Aramco Digital, SDAIA, and Humain represent. Where those programs are oriented toward economic productivity — extracting value from AI in oil and gas operations, government services, and commercial enterprise — NEOM is oriented toward demonstrating Saudi Arabia’s technological ambition to an international audience.
The showcase function is real and valuable for MBS’s Vision 2030 narrative. When international investors, heads of state, and technology executives visit Saudi Arabia, they visit NEOM. The impression they take away — of a country that is building something genuinely new, not just deploying existing technology — shapes perceptions of Saudi Arabia as an AI investment destination. That perception has commercial value even if NEOM’s AI systems never reach the scale originally advertised.
The practical AI contribution from NEOM to the broader Saudi ecosystem is different: it is a research environment. The operational data from NEOM’s energy management, transportation, and building systems will generate insights about AI-managed infrastructure that can be applied across Saudi Arabia’s conventional cities — Riyadh, Jeddah, Dammam — where the same efficiency gains are achievable through retrofitting. NEOM as a living laboratory for smart city AI is commercially valuable even at 5% of its announced scale.
Forward Outlook: What Survives the Scaling
Three things survive NEOM’s scaling down and remain genuinely relevant for the Saudi AI buildout.
First, the DataVolt facility at Oxagon. This is a conventional data center project at an unconventional scale, in a location with renewable energy advantages that are not replicated elsewhere in the kingdom. The 1.5 GW target is ambitious but not implausible for a ten-year development timeline, and initial phases are proceeding. DataVolt at Oxagon will be a meaningful component of Saudi Arabia’s AI compute geography regardless of how many people live in The Line.
Second, Tonomus as the operational technology developer for NEOM’s actual built environment. The technology requirements for operating a purpose-built city at any scale — energy management, transportation, security, resident services — are real and will generate genuine deployment experience. Tonomus is building AI systems at a scale that is large enough to be commercially significant and architecturally interesting enough to be technically valuable.
Third, NEOM’s role as a foreign investment showcase. The megaproject’s continued existence, even at reduced scope, maintains Saudi Arabia’s position as a destination for advanced technology investment. The indirect effect on the broader AI buildout — attracting international partners who want to be associated with NEOM’s technological ambition — has been undervalued in assessments that focus exclusively on NEOM’s own construction timeline.
For vendors and investors, the practical conclusion is to scope NEOM opportunities against current construction commitments, not announced visions. The DataVolt facility and Tonomus technology contracts are real and fundable. The residential and commercial technology systems for 9 million Line residents are not yet real, and the timeline for them to become real has extended significantly. The distinction between these two categories is the difference between commercially viable NEOM opportunities and commercially premature ones.
Power, Cooling, and the Renewable Energy Advantage
NEOM’s location in the Tabuk region of northwestern Saudi Arabia provides genuine, durable advantages for AI infrastructure that are independent of The Line’s scale trajectory. The region receives approximately 2,200 kilowatt-hours per square meter per year of solar energy — among the highest solar resource densities on earth — and the planned Neom Green Hydrogen project, a multi-billion-dollar facility being developed by ACWA Power, Air Products, and NEOM, will convert solar energy into hydrogen at industrial scale.
For data center operators, access to cheap renewable electricity is an increasingly critical site selection factor. AI compute infrastructure is energy-intensive: a single rack of 16 NVIDIA H100 GPUs draws approximately 10-14 kilowatts, and a modern hyperscale AI data center campus at 100 MW consumes more electricity annually than a small city. At 1.5 GW, DataVolt’s planned facility would require electricity procurement at a scale that makes the marginal cost of power one of the most important economic variables in the facility’s financial model.
Saudi Arabia’s renewable electricity costs — especially in high-solar-resource locations like Tabuk — are among the lowest in the world for new-build projects. The combination of extreme solar irradiance, available land, and government subsidized infrastructure development means that NEOM’s data center operators can access electricity at prices that comparable facilities in the US, EU, or East Asia cannot match. This economic advantage persists regardless of The Line’s architectural fate.
The Red Sea coast location also provides passive cooling advantages. Seawater cooling — drawing seawater from the Red Sea to cool data center systems through heat exchangers — can reduce cooling energy consumption by 30-40% compared to air-cooled facilities in hot climates. NEOM’s Oxagon industrial hub is sited on the coast specifically to enable maritime access for industrial operations, and that coastal positioning is directly valuable for data center cooling.
The NEOM Investment Decision Framework
Investors and vendors considering NEOM-related opportunities in 2025 and beyond benefit from a clear decision framework that separates the viable from the premature.
Construction-phase opportunities — civil engineering, power infrastructure, fiber networking, data center construction for the DataVolt facility — are proceeding at a reduced but real scale. Vendors who won contracts based on original NEOM scale assumptions should renegotiate or restructure those contracts against current scope; vendors entering new conversations should price against current commitments with optionality for scale expansion rather than assuming original scale materializes.
Tonomus technology contracts — for city management AI, resident services, building management systems — are viable for the scope of NEOM that will actually be built in the near term. Contracts sized for a 2.4-kilometer segment of The Line and a functioning Oxagon industrial hub are real opportunities. Contracts premised on 9 million residents and 170 kilometers of linear city should be discounted heavily on probability.
DataVolt long-term power and data center agreements are the highest-quality NEOM-adjacent opportunity. The renewable energy advantage and coastal cooling access make Oxagon a structurally competitive data center location regardless of NEOM’s residential ambitions. Data center operators who secure long-term power purchase agreements at NEOM’s renewable energy economics are acquiring a structural cost advantage that compounds over the facility’s operational life.