Hexagon Data Center: 480 MW and the Sovereign Cloud Architecture
The Hexagon Data Center is the largest single sovereign-cloud-eligible facility in Saudi Arabia’s current infrastructure pipeline, and its 480 megawatt specification positions it as a cornerstone of the Kingdom’s ambition to host not just AI compute but the full spectrum of sensitive government and enterprise workloads on in-country, sovereign-compliant infrastructure. At 480 MW, Hexagon is a facility of global significance: larger than most hyperscaler data centers worldwide, capable of hosting more AI compute than the largest public training clusters announced as recently as 2023, and designed for a market — Saudi Arabia’s sovereign cloud and AI compute sector — that is transforming from nascent to mature within a compressed multi-year window.
To put the 480 MW specification in competitive context: Hexagon is approximately 2.4 times the size of HUMAIN’s Riyadh Phase 1 campus (200 MW), more than 1.5 times the size of HUMAIN’s Dammam campus (300 MW planned), and 2.4 times the size of the GDH Riyadh campus (200 MW). Among Saudi facilities currently under construction or operational, only DataVolt’s planned 1.5 GW NEOM facility exceeds Hexagon’s ambition — and DataVolt targets 2028 delivery while Hexagon is already under active construction. For the period 2025-2027 when Saudi AI infrastructure is coming online rapidly, Hexagon will be the single largest operational carrier-neutral facility in the Kingdom.
Sovereign Cloud Certification: The Technical and Regulatory Framework
Hexagon’s SDAIA-aligned sovereign cloud designation is a commercial differentiator with direct revenue implications. Sovereign cloud certification means Hexagon meets or is on a path to meeting the technical, legal, and operational requirements for hosting Saudi government workloads under KSA-RoD data residency frameworks, PDPL compliance obligations, and national security data handling requirements.
The certification process is multi-dimensional. Physical security requirements include perimeter access controls meeting government standards, biometric access logging for all data hall entry, CCTV coverage with government-specified retention periods, and the ability to host government oversight inspections without operational disruption. Network architecture requirements include physical isolation of government tenant infrastructure from commercial workloads — separate switching fabrics, not merely VLANs — and controlled, audited interconnection points with Saudi government networks.
Personnel requirements are often the most operationally challenging dimension of sovereign cloud certification in Saudi Arabia. Government workloads may require that all staff with physical or logical access to hosting infrastructure are Saudi nationals with applicable security clearances. At Hexagon’s scale — 480 MW requires hundreds of operational staff — satisfying this requirement means building a large team of cleared Saudi nationals, which represents a significant recruitment and training investment over a multi-year period. This is also a Vision 2030 Saudization alignment: the sovereign cloud certification requirement creates economic incentive to invest in Saudi technical workforce development.
Audit and compliance requirements include maintaining audit logs of all access events, change management records, and incident reports in formats specified by Saudi regulatory bodies, with retention periods and custody chain documentation that allow reconstruction of the complete operational history of any data processing event. For enterprise customers using Hexagon for PDPL-regulated data processing, these audit records are the technical evidence that supports their own compliance demonstrations.
Carrier-Neutral Operations: The Interconnection Hub Model
Hexagon’s carrier-neutral operating model is its fundamental competitive positioning against HUMAIN-tied and hyperscaler-tied alternatives. Carrier neutrality means Hexagon has no preferred relationship with any specific cloud provider, network operator, or technology vendor — all tenants are treated equally, competitive cloud providers coexist, and the facility’s commercial incentive is to maximize tenant density rather than to favor any one tenant’s ecosystem.
This model creates specific value for hyperscalers entering the Saudi market. AWS, Microsoft Azure, and Google Cloud are simultaneously cooperating partners (they all need Hexagon’s facility space) and competing vendors for Saudi enterprise customers. Each hyperscaler is uncomfortable co-locating in a facility where its primary cloud competitor has an equity relationship or strategic preference. Hexagon’s neutral model eliminates this concern: every hyperscaler has the same access on the same terms, and Hexagon’s commercial success depends on attracting all of them, not any particular one.
The interconnection hub function that develops in a mature carrier-neutral facility creates network effects that compound over time. When AWS, Azure, and Google Cloud are all present in Hexagon, they can exchange traffic with each other over short fiber paths rather than external network connections — cross-connects that typically cost a few hundred dollars per month versus several dollars per megabit per month for transit. Enterprise customers co-located in Hexagon get direct, low-latency access to all major cloud platforms simultaneously. Saudi internet service providers and network operators peer at Hexagon to minimize transit costs and improve performance. This interconnection ecosystem, once established, creates significant switching costs for all participants and sustainable competitive advantage for Hexagon as the hub.
The UAE’s G42 comparison is instructive for understanding what Hexagon is not. G42, backed by Mubadala and with deep Microsoft partnership relationships, operates a mixed model that combines captive AI infrastructure with commercial data center assets. The Microsoft-G42 partnership created a multi-billion-dollar committed relationship that effectively aligns G42’s largest facility with Microsoft’s ecosystem. Hexagon’s pure carrier-neutral model deliberately avoids this kind of strategic alignment, trading the revenue certainty of a large anchor hyperscaler relationship for the broader market positioning that true neutrality provides.
AI Workload Capacity: Translating 480 MW to GPU Count
Translating Hexagon’s 480 MW capacity into AI workload terms requires power density assumptions that depend on the hardware mix and cooling technology deployed.
Modern AI training infrastructure — NVIDIA GB300 Grace Blackwell supercomputers, AMD MI300X servers — operates at 80-100 kW per rack in high-density liquid-cooled configurations. At 90 kW per rack average for AI training workloads and 480 MW total power, Hexagon can support approximately 5,333 AI training racks. Each rack housing a single GB300 supercomputer contains 72 Blackwell B200 GPUs, putting Hexagon’s maximum AI training configuration at approximately 384,000 B200-class GPUs — enough to train multiple frontier-class language models simultaneously with substantial headroom.
At more realistic blended configurations — where AI training racks at high density coexist with AI inference racks at moderate density, storage systems, and conventional enterprise compute at standard densities — the effective capacity is better characterized as hosting 150,000-250,000 AI accelerators across a mixed workload tenant population. This is still a globally significant figure: very few data center campuses globally can claim this AI accelerator density.
The cooling technology enabling these densities at Hexagon will be predominantly liquid cooling — either rear-door heat exchangers, direct liquid cooling (DLC) to processor cold plates, or immersion cooling for the highest-density installations. Air cooling becomes inadequate above approximately 15-20 kW per rack for modern GPU hardware; high-density AI infrastructure requires liquid cooling as a fundamental facility requirement. Saudi Arabia’s climate — hot and dry — makes air cooling for high-density AI compute even less viable than in temperate climates, reinforcing the liquid cooling requirement. Hexagon’s under-construction design incorporates the liquid cooling infrastructure required to support the full 480 MW capacity at AI-appropriate rack densities.
Riyadh Location: Capital City Advantages for Sovereign Cloud
Hexagon’s Riyadh location is optimal for its sovereign cloud and enterprise colocation market positioning. The Saudi government ministries, major Saudi banks, Saudi Aramco’s Riyadh operations, and the headquarters of Vision 2030’s major project developers are all Riyadh-based — and all are potential Hexagon tenants for PDPL-compliant enterprise workloads and government-adjacent AI programs.
The connectivity infrastructure anchored in Riyadh benefits Hexagon’s carrier-neutral hub ambitions. stc’s core national network, the primary international carrier exchange points, and the concentrated enterprise WAN infrastructure that large Saudi institutions use are all more accessible in Riyadh than in the Eastern Province. A 480 MW carrier-neutral facility in Riyadh can develop the carrier and cloud provider ecosystem that makes it a true interconnection hub — the Saudi equivalent of Equinix’s major exchange facilities in major global metros.
Saudi Arabia’s AI infrastructure buildout is creating a network effects competition between Hexagon, Center3, and GDH for the role of Riyadh’s premier carrier-neutral interconnection hub. The facility that establishes the richest carrier and cloud ecosystem first will develop compounding advantages — each new tenant makes the facility more valuable to all existing tenants. Hexagon’s scale (480 MW) and sovereign cloud certification give it structural advantages in winning large government and enterprise anchor tenants that accelerate this ecosystem development. Track Hexagon’s construction progress and competitive positioning at Infrastructure.
Financial Structure and Developer Profile
Hexagon Data Center’s development is backed by capital structures that reflect the institutional sophistication of the Saudi AI infrastructure market. The facility’s sovereign cloud positioning and carrier-neutral model attract a specific investor profile: infrastructure funds that value long-dated contracted revenues, PDPL-compliant hosting premium pricing, and the strategic position as Saudi Arabia’s flagship sovereign cloud infrastructure asset.
The development entity behind Hexagon has positioned the facility’s sovereign cloud credentials as the primary commercial differentiation — justifying pricing above commodity colocation rates for government and enterprise tenants who require the elevated compliance posture. Sovereign cloud certified data center capacity in Saudi Arabia is genuinely scarce: SDAIA’s captive infrastructure is not available commercially, HUMAIN’s AI factory has a national program mandate that limits commercial colocation, and Center3’s 100 MW operational capacity is insufficient for the total sovereign cloud demand. Hexagon’s 480 MW at full build represents the first large-scale commercially available sovereign cloud capacity in Saudi Arabia, which justifies premium pricing that supports the facility’s financial model.
For international enterprises establishing Saudi operations under PDPL requirements — banks, healthcare companies, technology platforms — Hexagon’s sovereign cloud certification provides the compliance certainty that makes Saudi market entry commercially viable. Without certified sovereign cloud infrastructure, PDPL compliance for international companies requires either building proprietary Saudi data centers (capital-intensive and operationally complex) or accepting compliance risk on non-certified commercial hosting. Hexagon eliminates this dilemma, and the commercial value of that solution supports the pricing and utilization rates that make the 480 MW facility financially viable.
Power Infrastructure at 480 MW: Dedicated Substation Requirements
The power infrastructure investment required for Hexagon’s 480 MW capacity is an often underestimated dimension of the facility’s total development cost and timeline. Connecting 480 MW to Riyadh’s electrical grid requires dedicated high-voltage infrastructure at a scale that qualifies as utility-level investment.
A typical 480 MW data center requires at least two independent 380kV/132kV or 132kV/33kV grid connections, each capable of supplying the full 480 MW load for full N+1 power redundancy. Each grid connection requires a dedicated grid-switching station, step-down transformers of 200+ MVA rating, medium-voltage distribution switchgear, and the protection, metering, and control systems that Saudi Electricity Company requires for large industrial connections. The total transformer investment for 480 MW of dual-redundant power supply is measured in hundreds of millions of riyals.
Large power transformers of the type required for Hexagon’s grid connections are subject to the same 12-24 month global supply constraint that affects all large data center projects. Hexagon’s project development timeline — likely requiring transformer orders placed 18-24 months before facility commissioning — means the power infrastructure supply chain is the longest-lead critical path item in the construction schedule. Any project delay or design change that affects the electrical infrastructure after transformers are ordered creates cost and schedule risk that is difficult to recover.
Hexagon and the NEOM/Oxagon Connectivity
Hexagon’s Riyadh location creates an important complement to the DataVolt-NEOM facility at Oxagon. While DataVolt targets net-zero AI compute at massive scale for sustainability-premium customers, Hexagon’s sovereign cloud positioning and carrier-neutral interconnection model serves the government and enterprise connectivity requirements that NEOM-adjacent infrastructure cannot address. The two facilities serve different market segments and potentially different geographic catchment areas, making them complementary rather than directly competitive within Saudi Arabia’s total infrastructure plan.
For customers that need both the regulatory compliance and interconnection of Hexagon’s Riyadh facility and access to net-zero computing at DataVolt-NEOM, the long-distance fiber connectivity between Riyadh and NEOM’s Tabuk region — being developed as part of Saudi Arabia’s Vision 2030 national fiber backbone — creates the technical foundation for hybrid deployments that use both facilities. This geographic diversification of AI compute infrastructure is itself a resilience asset: a natural disaster, power disruption, or security event affecting one region does not compromise all of Saudi Arabia’s AI compute capacity.