The True Cost of Building Hyperscale Capacity in the Kingdom
Saudi Arabia is in the middle of the largest national data-center construction program ever attempted by a single sovereign in a five-year window. The Public Investment Fund, Humain, Aramco Digital, Center3, the regional hyperscalers, and a layer of sovereign-friendly developers are collectively underwriting capacity that, when fully delivered, will exceed 6.5 gigawatts of IT load between Riyadh, the Eastern Province, and NEOM. The headline numbers are widely reported. The construction economics underneath them are not. This analysis lays out the typical 2026 range for $/MW capex on Saudi hyperscale builds, breaks the cost into shell, fit-out, and AI-ready overlays, and frames it against UAE and US benchmarks.
The Headline Number: $9.8M to $14.5M per MW for AI-Ready Tier IV
A fully delivered, AI-ready, Tier III+ to Tier IV hyperscale data center in Saudi Arabia in 2026 costs in a typical range of $9.8M to $14.5M per MW of IT load, all-in. That is a wide band, and the spread is real — it captures the difference between a fast-track Riyadh build on pre-permitted Modon land using a domestic EPC versus a NEOM-region build with bespoke architectural requirements, off-grid renewable integration, and full liquid-cooling readiness for B200/GB200 density.
This compares to US benchmarks of roughly $8.5M to $13M per MW for equivalent AI-ready hyperscale, and UAE benchmarks of roughly $9.5M to $13.5M per MW. Saudi sits at a slight premium to the US baseline driven by local-content requirements, longer logistics chains for specialist equipment, and the cooling overhead imposed by ambient design temperatures that routinely exceed 45°C in summer. The Saudi premium versus the US has compressed meaningfully since 2023, when comparable builds carried a 25 to 35 percent uplift; the maturing local supply chain and the volume effects of the Humain and PIF buildouts have pulled costs down.
Shell Construction: $2.8M to $4.6M per MW
The shell — the building, the foundation, the structural envelope, the basic utility connections — accounts for $2.8M to $4.6M per MW in the Kingdom in 2026. The lower bound reflects warm-shell, single-story, Riyadh-metro builds on flat industrial-zone land with pre-existing road, water, and grid connections. The upper bound reflects multi-story, blast-resistant, NEOM-region builds with bespoke seismic provisioning and remote-site logistics overhead.
Local-content requirements imposed by the Ministry of Investment and the Saudi Authority for Industrial Cities and Technology Zones (Modon) push a meaningful share of shell construction value to domestic contractors and material suppliers. In practice this means concrete, rebar, structural steel, and standard MEP components are sourced from Saudi suppliers (Hadeed, Al-Yamama Cement, the Tasnee polymer chain), which is broadly cost-neutral or slightly cheaper than imports. Specialist materials — high-spec switchgear, certain UPS modules, specific HVAC chillers — still import at a 5 to 12 percent landed-cost premium versus US or European benchmarks.
Fit-Out: $5.8M to $9.2M per MW for AI-Ready Capacity
The fit-out — power distribution, UPS, generators, cooling, fire suppression, structured cabling, security, and the AI-specific overlays — is where Saudi builds in 2026 diverge most sharply from legacy hyperscale benchmarks. AI-ready fit-out for B100/B200/GB200 density at 80 to 130 kW per rack costs in a typical range of $5.8M to $9.2M per MW of IT load, compared to roughly $4M to $6M per MW for legacy 8-15 kW per rack air-cooled fit-out.
The drivers are well understood: liquid cooling infrastructure (CDUs, manifolds, rear-door heat exchangers, in-rack distribution) adds $800K to $1.4M per MW; redundant 415V or 240V DC distribution at higher amperage adds $400K to $700K per MW; and the higher generator and UPS capacity needed to back AI-density loads adds another $300K to $600K per MW. Add specialist fire suppression for liquid-cooled environments and AI-grade structured cabling at 400GbE and 800GbE, and the AI-ready overlay alone runs $1.8M to $3.2M per MW above legacy fit-out.
EPC Counterparties and Pricing Power
The EPC market for Saudi hyperscale has consolidated around a roughly six-counterparty panel in 2026. Bechtel, AECOM, Black & Veatch, and Jacobs serve the international tier with full-program EPC capabilities. Saudi Binladin Group, Almabani General Contractors, Al-Rashid Trading and Contracting, Nesma, and Al-Khodari serve the domestic tier, often in joint venture with international engineering counterparties for the mission-critical scope. The Public Investment Fund-aligned developers increasingly favour JV structures that meet local-content thresholds while preserving international engineering rigor on the AI-critical scope.
Pricing power on the EPC panel has tightened noticeably during 2025-2026 as the volume of concurrent megawatt commissioning has stretched the available specialist labour pool. EPC margins on Saudi hyperscale in 2026 typically run 8 to 14 percent, compared to 5 to 9 percent in less constrained US and European markets. This is a real cost component for buyers and one of the structural reasons the Saudi premium versus US baseline has not fully closed.
NEOM-Specific Construction Costs
NEOM is its own pricing universe. Construction in the NEOM region — Oxagon, the Line corridor, the Sindalah-adjacent compute zones — runs at a typical premium of 18 to 32 percent above Riyadh-equivalent builds. The drivers are remote-site logistics, bespoke architectural requirements imposed by the NEOM master planning authority, and the integration of off-grid or partially-off-grid renewable power systems (NEOM Green Hydrogen, the NEOM Solar megaproject, on-site BESS).
A 100MW AI-ready hyperscale build in NEOM in 2026 costs in a typical range of $1.2B to $1.65B all-in, versus $980M to $1.45B for the same scope in the Riyadh metro. The premium buys integrated renewable power, exemption from certain regulatory overhead, and the marketing value of the NEOM brand for sovereign-friendly tenants. Whether it is worth the premium depends entirely on the tenant’s strategic positioning.
Local-Content Premiums and the Nitaqat Layer
The Saudi local-content framework, administered through the Local Content and Government Procurement Authority, requires hyperscale developers to meet thresholds for domestic spend across construction materials, specialist services, and Saudi-national employment. In practice this adds a typical premium of 3 to 7 percent on total construction cost versus an unconstrained import-driven build, but the premium is partially offset by tax incentives, accelerated permitting, and access to PIF-aligned co-investment.
Nitaqat employment requirements add another layer: the construction phase requires a minimum percentage of Saudi-national workers across skill categories, and meeting the requirement during peak commissioning often requires a 5 to 12 percent labour-cost uplift versus expatriate-only workforces. The Ministry of Human Resources and Social Development has signaled tightening of these requirements through 2027, which will keep upward pressure on construction labour costs.
Land, Permitting, and Soft Costs
Land and permitting are covered in detail in a separate analysis, but the construction-cost view should include them as soft-cost adjacencies. Saudi hyperscale builds in 2026 typically carry $280K to $720K per MW in land cost for Riyadh-metro Modon zones, with NEOM and Eastern Province sites varying widely. Permitting, environmental impact assessments, and Cloud SEZ designation costs add another $120K to $310K per MW on average.
The 2026-2028 Cost Trajectory
The trajectory through 2027-2028 points toward modest cost compression on legacy fit-out scope (driven by maturing domestic supply) but continued upward pressure on AI-ready overlays as B200 and Rubin-generation density requirements push liquid-cooling sophistication. Net-net, we expect the typical 2026 range of $9.8M to $14.5M per MW to migrate toward a 2028 range of roughly $9.5M to $15M per MW, with the spread widening rather than narrowing as bespoke AI architecture diverges further from commodity hyperscale.
What Buyers Should Negotiate
Buyers entering the Saudi hyperscale construction market in 2026 should treat the following as negotiable: EPC margin, contingency reserve, schedule float, change-order pricing methodology, and the split between domestic and international scope. Treat the following as effectively fixed: local-content thresholds, Nitaqat labour requirements, and the underlying material and equipment cost stack. The biggest savings opportunities sit in EPC structuring (single-prime versus multi-prime) and in the timing of long-lead-item procurement against commissioning milestones.
These are analytical estimates synthesised from public buildout activity, EPC panel positioning, and observed cost trends. Specific project economics vary based on site, scope, schedule, and counterparty mix, and these figures should not be treated as committed price quotes.
Long-Lead Equipment and Supply-Chain Dynamics
The single most schedule-critical category of construction-cost component in Saudi hyperscale builds in 2026 is long-lead equipment. Medium-voltage switchgear, large UPS modules, generator sets at 3MW+ capacity, and specialist HVAC chillers all carry lead times of 32-72 weeks from order to delivery, and the Saudi buildout volume during 2024-2026 has stretched this further. Operators with disciplined long-lead-item procurement programs — placing orders 18-24 months ahead of commissioning — typically capture pricing 8-14 percent below late-procurement equivalents and avoid the schedule-slip risk that has affected some of the more aggressive 2024-2025 starts.
The cost premium for expediting long-lead equipment in 2026 is meaningful. Standard 30MW generator sets with normal lead times price in a typical range of $3.6M to $5.2M per unit installed; expedited delivery (8-12 week compression) typically commands a 18-32 percent premium plus expedited freight cost. UPS modules at the 1MW+ tier follow a similar pattern. Buyers structuring their EPC contracts should explicitly address long-lead-item procurement responsibility — owner-furnished versus contractor-furnished — and the associated schedule and cost risk allocation.
Commissioning, Testing, and Cx Costs
The commissioning and testing phase — typically the final 12-20 weeks of a hyperscale build — adds $280K to $620K per MW of incremental cost across the various scopes (electrical, mechanical, fire-life-safety, IT-systems, integrated systems testing). The Saudi commissioning market has matured meaningfully during 2024-2026, with both international Cx specialists (Cxteam, Sibley, Henderson) and emerging Saudi-native Cx capability serving the buildout volume.
Quality-of-commissioning has direct operational cost implications. A facility commissioned to typical industry standards typically operates at design-PUE within 4-8 weeks of occupancy; a facility with poor commissioning may take 12-26 weeks to reach design-PUE and operates above design through that window. The cost of poor commissioning, expressed as excess operating cost during the ramp-to-design-PUE window on a 100MW hyperscale, is typically $2.4M to $5.8M — comfortably more than the cost premium for higher-quality Cx scope.
The 100MW Reference Project Total
For an integrated reference, a 100MW AI-ready Tier IV hyperscale build in the Riyadh metro on Modon-Sudair land in 2026 typically totals approximately $1.05B to $1.35B all-in, breaking down roughly as: $290M-$420M shell, $590M-$880M fit-out (including AI-ready overlay), $30M-$72M land and permitting, $28M-$62M long-lead procurement premium and expedite, $28M-$62M commissioning and testing, plus contingency of 6-10 percent. NEOM-region equivalents run roughly $1.2B to $1.65B all-in for the same IT-load scope.
Modular and Prefabricated Approaches
A growing minority of Saudi hyperscale builds in 2026 incorporate modular and prefabricated approaches — pre-engineered power-and-cooling modules, prefabricated data hall sections, factory-integrated MEP units. The capex impact is typically a 6-12 percent reduction versus stick-built equivalents at matched specification, with the larger benefit being schedule compression of 15-30 percent versus traditional construction. Vendors leading this category in the Saudi market include Vertiv (PowerMod modular UPS and switchgear), Schneider Electric (EcoStruxure modular DC), and the various prefabricated-data-hall specialists serving the GCC market. Schedule compression carries direct economic value as discussed in the cost-of-delay analysis, often making the modular approach attractive even at price parity with stick-built. Saudi tenants under aggressive timeline pressure (the Humain ramp, the SDAIA programs, certain hyperscaler region expansions) increasingly specify modular as the default approach for power and cooling scope.
Refurbishment and Brownfield Conversion
A small but growing subset of the Saudi DC market involves refurbishment of existing industrial facilities for digital-infrastructure use. Brownfield conversion economics typically run $5.8M to $9.4M per MW for AI-ready capacity, modestly cheaper than greenfield equivalents but with material constraints around floor-loading, ceiling height, and existing-grid-connection capacity. Brownfield is rarely the preferred option for hyperscale anchor builds but is a viable approach for mid-market and edge-class deployments where speed-to-market dominates the economics.
Financing Structures and Capex Funding
The capex funding structures supporting Saudi hyperscale construction in 2026 vary meaningfully across the buyer panel. PIF-aligned developers (Humain, the various PIF-associated DC operators) typically fund through a mix of PIF equity contribution and project-level debt arranged through the major Saudi banks (NCB, Riyad Bank, SABB) at sovereign-grade financing costs. Hyperscaler self-builds typically fund through corporate balance-sheet capital with internal-cost-of-capital pricing. Independent developers fund through sponsor equity, project-level debt, and increasingly through Islamic finance structures that have become competitive with conventional debt for Saudi-resident projects. The financing-cost component of total project economics typically runs 8-14 percent of total capex when expressed as cumulative interest-and-fee cost over the construction-and-commissioning window, with the lowest tier of this range available only to sovereign-grade and major-hyperscaler counterparties.
For deeper reading: