Walk into the Riyadh north industrial corridor in early 2026, drive past the King Abdullah Financial District towers, and within forty minutes you reach the Center3 Riyadh campus where stc’s data center business has been quietly stitching together the largest live capacity footprint in the Kingdom. The site hums in the way that big data centers always do: chiller plants on the perimeter, the constant low whine of computer-room air handlers, blue ground-fault indicators on outdoor switchgear, two new feeder substations commissioned in the second half of 2025, and behind the chain-link the unfinished steel of phase two — another seventy megawatts due to commission through 2026. Drive five hours east to Dammam and the picture sharpens further. There, on a piece of cleared desert near the Aramco eastern operations zone, the surveyors have already pegged out the perimeter of what Humain — the new sovereign AI company that the Public Investment Fund launched in May 2025 — is calling its first national AI training campus, with NVIDIA Blackwell racks scheduled for staged delivery through 2026 and 2027. This is the Saudi data center boom of 2026, and after twenty-four months of slow-cooked planning the megawatts are finally being poured into concrete.
The numbers tell the strategic story. Saudi Arabia ended 2024 with approximately 250 megawatts of installed live data center capacity across all operators. By 2026 the working figure is 400 to 500 megawatts, on a Vision 2030 trajectory toward roughly 1,300 megawatts by decade end — a build-out that implies five times the current footprint in six years. The capital flow underwriting that build-out is anchored on three names: ACWA Power, with a 1.5-gigawatt joint venture target with Saudi Aramco by 2032 and 5 to 7 billion US dollars committed; stc through its Center3 platform, expanding from roughly 150 megawatts in 2024 toward 350 megawatts by 2027 with a 4-billion-dollar-plus infrastructure programme; and Humain, the PIF sovereign AI company with an initial 40-billion-dollar commitment and a 1,000-megawatt national AI infrastructure target by 2030. Behind those three sit Mobily, Saudi Aramco Cloud, and a growing wave of foreign hyperscaler regions from Google, Microsoft, Oracle, and — eventually — Amazon Web Services.
This is a working analyst brief on the Saudi data center build-out for 2026 and the years immediately ahead. It covers the capacity numbers in megawatts and the Vision 2030 trajectory, the major Saudi player set with their capacity targets and capital commitments, the hyperscaler partnership economics, the power-cost arbitrage that anchors the entire investment thesis, the location preferences across Riyadh, Dammam, Jeddah, NEOM, and Yanbu, the comparison against UAE and Egyptian alternatives, the listed-equity and contractor investment vehicles available to foreign capital, the 2026 milestone calendar, and the risks that need to be priced before the cheque is signed. It is written for the infrastructure analyst, the cloud-services procurement lead, the Tadawul-curious foreign portfolio manager, and the corporate development executive trying to understand whether Saudi data center exposure deserves a real position rather than a footnote.
The Capacity Story: 250MW To 1,300MW On A Vision 2030 Trajectory
The capacity arithmetic is the right place to start because it frames every other decision. In 2024, Saudi Arabia’s total installed data center capacity was approximately 250 megawatts of IT load — measured the way the industry measures it, which is the power-to-rack capacity available to colocation customers, not the gross facility power that includes cooling and other ancillaries. Of that 250 megawatts, stc Center3 accounted for roughly 150 megawatts split between Riyadh, Jeddah, and a smaller Dammam footprint, making it by an order of magnitude the largest single operator. Mobily contributed roughly 50 to 80 megawatts. Saudi Aramco’s cloud arm accounted for a further 20 to 30 megawatts of energy-sector-focused capacity. Smaller operators and government in-house facilities filled the balance.
The 2026 working number — used by Saudi government planners, by Cushman & Wakefield’s MENA data center tracker, by Knight Frank’s regional infrastructure team, and by the major engineering consultancies producing capacity forecasts — sits at 400 to 500 megawatts of live IT load. The increment over 2024 reflects stc’s phase 2 commissioning at the Riyadh and Jeddah sites through 2025 and into early 2026, the first Humain phase coming online during 2026, plus incremental capacity from Mobily and the smaller operators.
The Vision 2030 target is approximately 1,300 megawatts by decade end. That number is built out of the operator-level commitments: 350 megawatts at stc Center3 by 2027 with further expansion through 2030; 1,000 megawatts at Humain by 2030; up to 1.5 gigawatts at the ACWA Power and Saudi Aramco joint venture by 2032 (with most of that beyond the 2030 horizon); and roughly 100 to 150 megawatts cumulative across Mobily, Saudi Aramco Cloud, and the smaller players. Net of operator overlaps and reasonable build-rate haircuts, 1,300 megawatts is the central case planning figure. Reuters technology coverage through 2025 and into 2026 has tracked the announcement and commissioning cadence in detail, with consistent reporting on Humain, ACWA Power, and the Center3 expansion.
For comparison, the UAE — the GCC’s other major data center hub through Abu Dhabi’s G42, Khazna Data Centres, and the Etisalat-backed footprint — is projected at roughly 800 megawatts by 2030 on industry forecasts. Egypt is projected at 200 to 300 megawatts. Qatar is smaller at roughly 100 to 150 megawatts. Israel sits at roughly 200 megawatts. The Saudi 1,300-megawatt target therefore positions the Kingdom as the largest single national data center market in the broader Middle East and North Africa region by a meaningful margin — a structural reordering of regional cloud and AI infrastructure that has implications for hyperscaler commercial strategy, Arabic-language AI training infrastructure, and government workload residency well beyond Saudi Arabia’s own borders.
Humain: The Sovereign AI Champion Model
The single most important new structural feature of the 2026 Saudi data center landscape is Humain, the sovereign AI company that the Public Investment Fund launched in May 2025. Humain is to Saudi AI infrastructure what Saudi Aramco is to Saudi hydrocarbons or what stc is to Saudi telecommunications: a national-champion vehicle wholly owned by the state, capitalised at sovereign scale, mandated to build a defining national capability across an entire industrial layer. The initial PIF commitment exceeded 40 billion US dollars at launch, with a strategic plan that anchors on 1,000 megawatts of data center capacity by 2030, primarily distributed between Riyadh and Dammam, with NEOM as a future expansion site.
The chip strategy is multi-vendor by design. Humain has signed framework agreements with NVIDIA covering supply of Blackwell-generation GPUs at scale, with phased delivery aligned to data center site readiness through 2026 and 2027. Parallel framework agreements with AMD cover Instinct MI-class accelerators, providing platform diversification against single-vendor exposure and giving Humain a dual-stack training capability matched against the architectural preferences of different end customers. The compute layer is wrapped in cloud and AI-as-a-service offerings positioned for Saudi government workloads, regulated industries, and enterprise customers across the GCC.
The customer book is being built deliberately. Humain’s announced relationships include Google — through cloud partnership rather than direct chip supply — Saudi Aramco for energy-sector AI, the Saudi government across multiple ministries through SDAIA workload coordination, and a growing list of regulated-industry buyers. Our companion piece on Humain’s Saudi AI 2026 strategy and deals walks through the specific commercial relationships and the timeline of public announcements. The Vision 2030 mandate is explicit: Humain is to be the operating-layer national AI infrastructure provider on which Saudi government, regulated, and strategic-industry workloads run, with foreign hyperscalers integrated as partners and capacity providers rather than as end-state operators of national-critical AI workloads.
For foreign investors, the Humain story matters even though Humain itself is private. The 40-billion-dollar PIF commitment cascades into procurement flows that benefit listed companies — NVIDIA and AMD on the chip side, the EPC contractors building the physical footprint, the equipment specialists supplying HVAC, electrical, and physical-security infrastructure, the connectivity and fibre operators wiring the campuses to subsea cables and to the hyperscaler regions, and the Tadawul-listed power and telecom companies that intersect with Humain’s value chain at multiple points. Our overview of the broader Saudi PIF portfolio holdings for 2026 contextualises Humain inside the wider PIF capital allocation strategy and shows where the data center thesis sits within the sovereign fund’s global portfolio.
stc Center3: The Established Operator Doubling Down
stc — Saudi Telecom Company, Tadawul ticker 7010 — is the established incumbent and the largest single data center operator in the Kingdom by live megawatts. The data center business was rebranded as Center3 in 2024 to separate the colocation and cloud-infrastructure offering from the consumer-telecom brand and to make the asset cleaner for hyperscaler partnerships and potential future strategic actions including a possible carve-out or partial listing. Center3 ended 2024 at approximately 150 megawatts of live IT capacity distributed across the Riyadh primary campus, the Jeddah secondary site, and a smaller Dammam footprint.
The committed expansion path takes Center3 to 350 megawatts by 2027 with capital expenditure of approximately 4 billion US dollars across the multi-year programme. The Riyadh expansion accounts for the largest single share, with phase 2 commissioning through 2025 and 2026 followed by phase 3 land already secured. Jeddah’s expansion is more modest in absolute megawatts but strategically important because Jeddah is the natural cable-landing site for Red Sea subsea fibre and therefore the preferred Saudi region for hyperscaler customer-facing inference latency-sensitive workloads. Dammam’s expansion supports Aramco-adjacent industrial customers and the Eastern Province workload base.
The customer book is anchored on Saudi government cloud workloads. Saudi cloud-residency rules — coordinated by SDAIA, the Saudi Data and AI Authority, and CITC, the Communications, Space, and Technology Commission — require regulated-data and government workloads to run inside the Kingdom. That requirement is the structural anchor for Center3’s economics: it gives Center3 a defended government anchor demand book that is unlikely to migrate offshore on price competition. Hyperscaler customers — Google Cloud, Microsoft Azure, Oracle Cloud — typically operate as colocation tenants at Center3 sites for their Saudi cloud regions rather than building dedicated greenfield infrastructure, sharing the operational underlay while running their own customer-facing service layer above. Bloomberg’s technology desk has tracked the Center3 expansion in multiple longer pieces across 2025, focusing on the hyperscaler partnership economics and the broader Tadawul telecom investability story.
ACWA Power And Saudi Aramco: The Green Data Center Joint Venture
ACWA Power — Tadawul ticker 2082 — is the structurally most interesting equity exposure to the Saudi data center build-out for foreign investors who want a listed wrapper rather than direct project participation. ACWA’s traditional business is independent power and water production with a global footprint across Saudi Arabia, the broader GCC, Egypt, Morocco, South Africa, and Central Asia, anchored on a fleet of utility-scale solar, wind, gas, and water-desalination assets. The data center pivot, announced in 2024 and progressing through 2025 and 2026, leverages ACWA’s core competence on integrated power-plus-infrastructure delivery into the AI-data-center build-out.
The headline structure is a joint venture with Saudi Aramco and additional partners targeting up to 1.5 gigawatts of data center capacity by 2032, with capital commitments in the 5-to-7-billion-dollar range across the joint venture. The strategic anchor is the integration of ACWA’s renewable power generation pipeline — including the Sudair 1.5-gigawatt solar facility, several other utility-scale solar developments under PIF’s national renewables programme, and the broader green hydrogen platform — with adjacent data center capacity. The pitch is straightforward: Saudi Arabia is building 30-plus gigawatts of new solar and wind capacity by 2030. Some of that capacity should be co-located with data center load to capture the green-power story for hyperscaler ESG positioning, to lock in long-term low-cost power for AI training, and to monetise grid-edge generation that would otherwise compete for Saudi grid dispatch.
The first identified site sits at Yanbu on the Red Sea coast, with land allocation and preliminary engineering completed through 2025 and ground-breaking targeted for the fourth quarter of 2026. Yanbu’s logic is industrial — proximity to the Yanbu Industrial City heavy-industry cluster, an existing power generation footprint that simplifies grid interconnection, Red Sea cable-landing access for subsea fibre, and reduced competition with the Riyadh and Eastern Province sites for skilled construction labour. Additional sites are under consideration along the Red Sea coast and in the Eastern Province adjacent to Aramco operations. The Financial Times Middle East has run multiple pieces tracking the ACWA-Aramco joint venture since 2024, with detailed coverage of the green-hydrogen-plus-data-center thesis and the contractor selection process.
For listed-equity investors, ACWA Power is a more diversified vehicle than a pure data center operator. The data center contribution is one of multiple growth pillars layered on top of the existing power and water franchise, alongside the Vision 2030 renewables build-out and the international project pipeline. That diversification cuts both ways: it limits the single-name concentration of the data center thesis, and it dilutes the upside if the data center business turns out to be the highest-growth pillar of the next decade. Investors should size ACWA exposure on the basis of the integrated business rather than as a pure data center play.
The Smaller Players: Mobily, Saudi Aramco Cloud, And The Foreign Hyperscalers
Mobily — Etihad Etisalat, Tadawul ticker 7020 — is the smaller listed Saudi telecom and the more modest data center diversifier. Mobily’s data center footprint sits at roughly 50 to 80 megawatts, primarily in Riyadh, with incremental expansion targeted through 2027 and 2028. The strategic logic is similar to stc’s: telecom diversification away from connectivity revenue toward higher-multiple infrastructure-as-a-service revenue, anchored on Saudi residency-rule government and regulated workloads. Mobily’s relative scale disadvantage versus Center3 means it operates as a credible secondary supplier rather than the primary anchor; it also makes Mobily a more interesting M&A or carve-out candidate over the medium term as the Saudi data center landscape consolidates.
Saudi Aramco Cloud is the cloud-services arm carved out of Saudi Aramco proper, focused on energy-sector workloads and adjacent industrial verticals. The capacity footprint is smaller — 20 to 30 megawatts of dedicated capacity in 2024, expanding through 2026 — but the customer book is uniquely valuable because it sits inside the Aramco Group operational stack with privileged access to upstream, midstream, and downstream IT modernisation budgets. Aramco Cloud has acquired and partnered with TONOMUS — the NEOM AI subsidiary — and Atos for technology integration, sitting alongside Aramco’s broader digital transformation programme. For foreign investors, Aramco Cloud is not directly investable; it sits inside the Aramco Group structure. Indirect exposure runs through Saudi Aramco itself (Tadawul ticker 2222), and our Saudi Aramco versus ExxonMobil 2026 comparison covers the broader Aramco investment case where the cloud arm is one of multiple growth pillars layered on top of the core hydrocarbon business.
Foreign hyperscalers entering Saudi Arabia are the second-order growth flywheel for the entire build-out. Google Cloud announced the Saudi region in 2024 with operational launch targeted for 2026 and general availability in 2027, anchored on a strategic partnership with Aramco that provides the physical hosting layer and the customer relationships in the Saudi enterprise market. Microsoft Azure announced Saudi data zones in 2025 with phased general availability through 2026 and 2027, integrated with Center3’s hosting infrastructure. Oracle Cloud committed to a Saudi region with general availability in 2025, also primarily through Center3 hosting. AWS has held discussions with Saudi authorities and operators since 2023 but has not yet committed to a Saudi region as of early 2026 — the holdouts most often cited are commercial terms and the operating-control architecture for sovereign-data customers.
The hyperscaler pattern is consistent: a foreign cloud provider partners with a Saudi local operator (typically Center3, occasionally Humain, for some workloads ACWA), the Saudi partner provides the physical and connectivity infrastructure, and the foreign hyperscaler provides the platform-layer services, the customer relationships, and the global commercial muscle. Saudi residency rules require this partnership pattern for regulated and government data, which converts what could be a pure-foreign cloud regional expansion into a Saudi industrial-capacity build-out with foreign partners. NVIDIA, by contrast, supplies chips into Humain and Center3 but has not announced a Saudi data center build of its own — the chip-supplier role is straightforward and the build-out role is left to the Saudi sovereign and corporate operators. Our piece on NVIDIA’s Saudi Arabia 2026 chips and data center positioning walks through the chip-supply economics in more detail.
The Power-Cost Arbitrage: 3 To 5 Cents Versus 8 To 12 Cents
The single most important number in the entire Saudi data center investment thesis is the power cost. Saudi industrial electricity for licensed data center operators is priced at approximately 3 to 5 US cents per kilowatt-hour in 2026, against 8 to 12 cents for typical US data center hubs in Texas (Dallas-Fort Worth, Houston) and Virginia (Loudoun County’s Data Center Alley), and 15 to 20 cents in Singapore. That gap is structurally reproducible across the medium term because Saudi electricity is anchored on hydrocarbon feedstock at near-marginal-cost pricing, supplemented by an aggressive solar pipeline that delivers some of the lowest unsubsidised levelised cost of energy globally — in the range of 1.5 to 2.5 cents per kilowatt-hour for utility-scale solar contracts awarded under the Saudi National Renewable Energy Programme.
The arithmetic at hyperscale matters. A 100-megawatt AI training campus running at 80 percent annual utilisation consumes approximately 700 gigawatt-hours per year of electricity. The cost differential between Saudi pricing at 4 cents per kilowatt-hour and US Texas pricing at 9 cents per kilowatt-hour on that volume is roughly 35 million dollars per year. On a 500-megawatt campus — which is the scale Humain and ACWA-Aramco are planning — the differential rises to 175 million dollars per year. Over a ten-year operational window, the cumulative power-cost advantage runs to 1.75 billion dollars per gigawatt of capacity. That is large enough to justify the latency penalty on customer-facing inference, the training-data-residency adjustments, the regulatory friction of operating under Saudi data sovereignty rules, and the geopolitical premium that some customers may demand on Middle East infrastructure exposure.
The structural durability of the power-cost advantage is the core reason hyperscalers are committing to Saudi regions despite the latency disadvantage versus closer EU data centers and the customer-base friction versus US sites. AI training workloads are largely insensitive to network latency — the work happens inside the data center across a tightly coupled GPU cluster, and the customer-facing latency cost is incurred only at job-submission and result-retrieval time. Inference workloads are more latency-sensitive and therefore tend to remain closer to customers, but bulk training, fine-tuning, and embedding-generation workloads are perfectly suited to compute-heavy, latency-tolerant locations with cheap power. Saudi Arabia is offering exactly that profile.
The renewables overlay strengthens the case further. Saudi Arabia’s solar and wind pipeline by 2030 exceeds 30 gigawatts of new capacity. ACWA Power’s Sudair facility alone is 1.5 gigawatts of solar that will feed adjacent data centers under co-located power purchase agreements. The NEOM Helios green hydrogen project is being structured to feed the NEOM data center pilot with green-hydrogen-derived power for ESG-sensitive customers willing to pay a green-premium tariff. The combination of cheap baseline grid power plus an adjacent green-power overlay gives Saudi data center operators a hyperscaler ESG pitch that few alternative locations can match at comparable cost. The US Energy Information Administration’s Saudi Arabia country page provides structured macro data on Saudi power generation and the ongoing transition that anchors this investment thesis.
Location Map: Riyadh, Dammam, Jeddah, NEOM, Yanbu
The Saudi data center footprint is concentrated across five primary geographic clusters, each with distinct strategic logics that matter for capacity planning, customer base, and risk concentration.
Riyadh (Diriyah, KAFD, Riyadh Tech Valley)
Riyadh is the dominant cluster by live capacity and by 2030 trajectory. The capital region accounts for approximately 60 percent of current installed capacity and is forecast to retain roughly half of the 1,300-megawatt 2030 footprint. The strategic logic is simple: Riyadh is the population and economic centre, the home of every Saudi government ministry whose workloads sit at the centre of the residency-rule demand book, the King Abdullah Financial District financial-services cluster, and the headquarters footprint of essentially every major Saudi corporate. Customer-proximity matters even for AI-heavy workloads when latency-sensitive enterprise applications are layered on top, and Riyadh is the natural anchor. The major Riyadh sites include the Center3 north Riyadh primary campus, the Humain phase-1 Riyadh site, several Mobily smaller sites, and the planned hyperscaler customer-facing zones for Google, Microsoft, and Oracle.
Dammam And The Eastern Province
Dammam and the broader Eastern Province cluster — including Khobar and Dhahran — are the second cluster by 2030 capacity. The economics are anchored on cheaper land, abundant industrial power, and Aramco proximity for the energy-sector cloud workloads. Humain has identified the Eastern Province as the home of its second major training campus, with land secured and ground-breaking scheduled for 2026. Saudi Aramco Cloud’s footprint sits primarily in this cluster. Center3’s Dammam expansion is more modest in absolute megawatts but strategically important for Eastern Province customer proximity. The Eastern Province also benefits from its proximity to the Bahrain-Saudi Arabia subsea cable landings and to the Persian Gulf cable infrastructure that connects to broader regional connectivity.
Jeddah And The Red Sea Cluster
Jeddah is the Red Sea cluster anchor, with cable-landing advantages on the Red Sea subsea fibre routes that connect Saudi Arabia to Egypt, Sudan, and onward to Europe and India. Jeddah’s data center footprint is smaller than Riyadh and Dammam in absolute megawatts but strategically important for hyperscaler customer-facing regions because the cable-landing topology gives Jeddah-located capacity better latency to European and African customers than equivalent capacity in Riyadh. Center3’s Jeddah expansion is the major driver, with smaller incremental capacity from Mobily and from the planned hyperscaler regions.
NEOM
NEOM — the Vision 2030 mega-city under construction in the Tabuk region of north-western Saudi Arabia — is the future-facing data center cluster anchored on the green-hydrogen-derived power thesis. The NEOM Helios green hydrogen project will feed the NEOM data center pilot with carbon-free power, providing a premium ESG offering for hyperscaler customers willing to pay a green-tariff premium. NEOM’s near-term capacity is modest because the city is still under construction, but the medium-term plan is for several hundred megawatts of dedicated capacity by 2030 with growth into the 2030s. Our piece on the NEOM investment scorecard for Q2 2026 covers the broader NEOM build-out where the data center component is one growth pillar among several.
Yanbu
Yanbu, on the central Red Sea coast, is the ACWA Power-led industrial data center cluster, anchored on the joint venture with Saudi Aramco and aligned with the existing Yanbu Industrial City heavy-industry footprint. The strategic logic is integrated power-plus-infrastructure delivery using ACWA’s existing power generation assets, plus reduced labour and land cost competition versus the Riyadh and Eastern Province clusters. Yanbu’s data center development is at the early ground-breaking phase as of 2026 with operational capacity scheduled to come online from 2028 onward.
The 2026 Milestone Calendar
The investor monitoring stack for 2026 should be built around a specific milestone calendar that drives both equity moves and pipeline visibility for the next phase of capex.
Q2 2026: stc Center3 phase 2 commissioning. The Riyadh primary campus phase 2 — approximately 70 additional megawatts — is scheduled for live commissioning during the second quarter, with corresponding revenue ramp through the second half of 2026. The phase 2 customer book is largely pre-signed across hyperscaler colocation tenants and Saudi government workloads.
Q3 2026: Humain phase 1 going operational. The first Humain Riyadh data center campus — approximately 100 to 150 megawatts of initial IT load — is scheduled to come online during the third quarter, with NVIDIA Blackwell racks commissioned in staged tranches and AMD Instinct accelerators following on a parallel deployment timeline. Humain’s first major customer announcements are expected during this quarter.
Q4 2026: ACWA Power-Aramco data center joint venture ground-breaking at Yanbu. Physical construction commencement marks the formal capital-deployment phase of the joint venture, triggering EPC contractor mobilisation and the first phase of vendor-equipment procurement. Capital expenditure recognition begins in earnest from this point.
2027: Google Cloud Saudi region general availability. The general availability launch of the Google Cloud Saudi region marks the first major hyperscaler region operational at full commercial scale in the Kingdom, with corresponding revenue recognition and customer-base build-out. Microsoft Azure and Oracle follow on broadly similar timelines.
2028: NEOM data center first phase. The NEOM data center pilot first phase — anchored on green-hydrogen-derived power — comes online with initial capacity in the 50-to-100-megawatt range. Operational ramp follows through 2029 and 2030 as the broader NEOM city construction progresses.
Saudi Versus UAE Versus Egypt: The Honest Regional Comparison
Foreign capital looking at MENA data center exposure typically considers the trade-offs across Saudi Arabia, the UAE, and Egypt. The honest comparison matters because each market has genuine strengths and the right answer depends on the specific thesis.
The UAE — through Abu Dhabi-based G42 and its subsidiary Khazna Data Centres, and the Etisalat-backed footprint — has a head start on operational maturity, a deeper bench of hyperscaler relationships built up through a longer commercial cycle, and an earlier mover advantage on Arabic-language AI training infrastructure through G42’s collaboration with Microsoft and OpenAI. The UAE’s 2030 capacity is projected at roughly 800 megawatts. UAE strengths include the operating-ecosystem maturity, the international financial-services footprint that brings global enterprise customers, and the regulatory predictability of operating under DIFC and ADGM common-law frameworks.
Saudi Arabia counters with cheaper power (3 to 5 cents versus UAE’s 6 to 9 cents), larger land availability, government anchor demand at scale through the Vision 2030 ministerial workload book, and the explicit Humain national-champion model that the UAE does not exactly replicate. Saudi 2030 capacity at 1,300 megawatts is roughly 60 percent larger than the UAE’s projected 800 megawatts, reflecting the larger underlying economy and the more aggressive sovereign capital deployment.
Egypt is a smaller market — 200 to 300 megawatts projected by 2030 — focused on Egyptian and African workloads with cable-landing advantages on the Mediterranean and on the Red Sea. Egyptian power costs are competitive with Saudi Arabia at 6 to 8 cents per kilowatt-hour for industrial users, and labour costs are dramatically lower. Egypt’s structural disadvantage is a smaller domestic enterprise customer base and a less mature sovereign-capital deployment, but the cable-landing position makes Egypt strategically interesting for Africa-facing capacity.
The honest answer is that the Middle East and North Africa region will end the decade with two or three durable data center hubs operating in parallel rather than one winner. Saudi Arabia will be the largest by megawatts; the UAE will be the most operationally mature; Egypt will be the regional cable connectivity hub for African workloads. Foreign investors with regional theses should expect to operate across multiple jurisdictions rather than to pick one winner.
The comparison against US AI data center hubs in Texas and Virginia matters too. Saudi advantages are cheaper power, abundant land, and government-coordinated infrastructure delivery. US advantages are deep tech talent, customer base size, capital markets sophistication, and the regulatory framework that ultimately determines AI training-data access for cutting-edge models. The Saudi pitch to hyperscalers is AI training compute plus green power plus government anchor demand; it does not aim to replace US sites for the most cutting-edge work but rather to capture the bulk-training, fine-tuning, embedding-generation, and Arabic-language workloads where Saudi’s cost advantage and residency-rule lock-in are most powerful. The Wall Street Journal’s Middle East coverage has run multiple analytical pieces in 2025 and 2026 working through the Saudi-versus-US trade-off in detail.
Investment Vehicles For Foreign Capital
The investment-vehicle question is where the analytical work translates into actual portfolio positions. Foreign capital looking at Saudi data center exposure has five practical routes.
Direct listed equity on Tadawul. ACWA Power (ticker 2082) is the most direct integrated power-and-infrastructure exposure with the data center build-out as a meaningful growth pillar. stc (ticker 7010) is the established telecom incumbent with Center3 as the largest operating data center business. Mobily (ticker 7020) is the smaller telecom diversifier. All three are listed on Tadawul and accessible to foreign investors through the Qualified Foreign Investor framework or through the Saudi Exchange Direct Trading route.
Sovereign sukuk and quasi-sovereign debt. Humain is private but its capital base sits inside the broader PIF balance sheet. Foreign investors can access PIF-related cashflows indirectly through PIF sovereign sukuk issuance, which has been an active issuance programme through 2024 and 2025 with consistent foreign investor participation. Saudi Aramco’s quasi-sovereign debt instruments provide partial exposure to Aramco Cloud through the parent’s broader balance sheet.
Cloud services contracts and procurement positioning. Foreign technology vendors selling into the Saudi data center build-out — chip suppliers, equipment suppliers, EPC contractors, software platform providers — capture Saudi capital deployment into the procurement chain. NVIDIA and AMD are the two largest publicly listed beneficiaries on the chip side. The major server OEMs — Dell, HPE, Supermicro — capture a meaningful share of the equipment procurement. The networking specialists — Cisco, Arista — capture the connectivity layer. The HVAC and electrical specialists — Schneider Electric, Vertiv — capture the physical infrastructure layer.
Engineering and construction contractors. The EPC layer building the physical data center footprint includes a mix of international and Saudi contractors. Internationally listed beneficiaries include Bechtel (private), Hill International, AECOM, and the major European contractors. Saudi-listed contractors with data center exposure include several Tadawul-traded construction names that have won meaningful contracts on the Vision 2030 mega-project pipeline.
Hyperscaler equity exposure. Indirect exposure to Saudi data center growth runs through Google’s parent Alphabet, Microsoft, Oracle, and (eventually) Amazon. Saudi capacity is a small share of total global hyperscaler revenue, so this is exposure-by-correlation rather than direct exposure, but it captures the overall AI infrastructure thesis with broader diversification across global cloud demand. For listed-equity investors who want general AI infrastructure exposure with a Saudi tailwind included, the hyperscaler names provide the cleanest large-cap wrapper.
The Risk Stack: What Has To Be Priced Honestly
Saudi data center exposure carries genuine risks that any disciplined investment committee will press on. Five risks dominate the priced-risk discussion.
Hyperscaler concentration risk. Google, Microsoft, and Oracle dominate the announced Saudi region commitments. A delay or cancellation by any one of the three would compress the demand pipeline materially. The risk is partially mitigated by Saudi government anchor demand through the residency-rule book, by Humain’s sovereign demand role, and by the diversification across multiple hyperscalers, but single-name hyperscaler delays would still hurt operator revenue ramp.
US export controls on advanced chips. The US Bureau of Industry and Security has tightened rules on advanced AI chip exports to Saudi Arabia at multiple points since 2023, citing concerns about Chinese chip diversion through Saudi intermediaries. The current 2026 framework permits Blackwell-generation NVIDIA shipments into Saudi data centers under licensed conditions, but the rules are subject to political volatility under shifting US administrations. A meaningful tightening would compress Humain’s compute-build trajectory and slow the broader build-out.
Construction cost inflation. Saudi Arabia’s Vision 2030 mega-project pipeline has crowded in EPC contractors, pushed up labour and material rates, and lengthened delivery timelines on data center projects. Construction-cost overruns of 15 to 30 percent versus original budgets are realistic baseline assumptions for projects approved in 2024 and 2025. Investors should model project IRR with construction-cost sensitivity rather than treating budget figures as fixed.
Power transmission grid bottlenecks. The Saudi grid is robust on generation but has localised transmission constraints, particularly in the Eastern Province around the Aramco operating zone and around Riyadh. Live commissioning of new data center capacity has been delayed in some cases by transmission-substation lead times even when generation capacity exists. Saudi Electricity Company has accelerated transmission investment but the constraint remains real for 2026 and 2027.
Talent shortage. Advanced AI infrastructure operations — site reliability engineering, custom-silicon networking, distributed training operations — require skill sets that Saudi Arabia is still building. The near-term solution is expatriate hiring through Humain and Center3 plus aggressive training programmes through KAUST and the Saudi technical university system. The medium-term gap is likely to be filled, but for 2026 and 2027 the talent constraint adds operational friction that must be priced.
The 2026-2030 Investment Thesis
The structural case for Saudi data center exposure rests on five forces that are unlikely to reverse over the medium term.
Power-cost arbitrage durability. Saudi industrial electricity will remain cheaper than US, European, and East Asian alternatives through 2030 and beyond, anchored on hydrocarbon feedstock economics plus a vast solar pipeline at internationally competitive levelised costs. The 35-to-50-million-dollar-per-100-megawatt-per-year cost advantage is reproducible across the medium term and is the single most powerful driver of hyperscaler commitments.
Vision 2030 sovereign capital deployment. The PIF, through Humain and through the broader Vision 2030 portfolio, is committing capital at sovereign scale into AI infrastructure. The 40-billion-dollar Humain commitment alone exceeds the global cumulative capital deployment of any private AI infrastructure operator. That capital base is patient, anchored on a 2030 strategic horizon, and structurally insensitive to the cost-of-capital cycle that affects pure-private operators.
Government and regulated demand book. Saudi cloud-residency rules through SDAIA and CITC create a defended demand book for in-Kingdom data center capacity that is unlikely to migrate offshore on price competition. That anchor demand provides the revenue floor that underwrites the broader build-out, separating the Saudi market from speculative greenfield builds in markets without comparable residency frameworks.
Hyperscaler regional commitments. Google, Microsoft, Oracle, and likely Amazon will operate Saudi cloud regions through 2030 and beyond. Once a hyperscaler region is live, the customer-acquisition and retention dynamics make exit costly, anchoring durable demand for the underlying physical capacity. The physical-capacity layer underneath the hyperscaler regions is largely operated by Center3, Humain, and ACWA, not by the hyperscalers themselves — capturing the value at the operator layer is the structural advantage of the Saudi build-out.
Arabic-language AI training infrastructure. The world’s roughly 400 million Arabic speakers represent a meaningful underserved market for advanced AI services, and Saudi Arabia is positioning itself — through Humain and through the broader sovereign AI mandate — as the primary infrastructure base for Arabic-language model training. That is a distinct workload pool from English-language AI training, with different data-residency, regulatory, and customer-base characteristics, and Saudi Arabia is structurally positioned to capture the bulk of that workload through 2030.
Practical Conclusions: Five Takeaways For 2026 Investors
First, take Saudi data center capacity numbers seriously. The 250-megawatt-to-1,300-megawatt build-out from 2024 to 2030 is real, the capital underwriting it is real, and the customer demand book is real. This is not a speculative thesis; it is a structural infrastructure cycle with sovereign anchoring.
Second, do not over-rotate to a single name. ACWA Power, stc, and Mobily each capture a different slice of the build-out. Humain is private. Hyperscaler exposure is global with a Saudi sliver. The right portfolio position is a basket across the available vehicles rather than a concentrated single-name bet.
Third, model the construction-cost and timing risk explicitly. Vision 2030 mega-project pipeline congestion creates real timing slippage and cost overruns. Project IRR estimates should be sensitivity-tested with 15-to-30-percent construction-cost overrun scenarios and 12-to-18-month delivery delays.
Fourth, monitor the milestone calendar quarter-by-quarter. The 2026 milestone book is dense — Center3 phase 2, Humain phase 1, ACWA-Aramco ground-breaking — and the equity reaction to milestone delivery (or slippage) is likely to be meaningful at each step. Active monitoring through Reuters, Bloomberg, and FT regional coverage is the operational requirement.
Fifth, treat Saudi data center exposure as part of a broader regional theme. The thesis works alongside UAE G42 exposure, alongside hyperscaler global positions, and alongside the broader AI infrastructure equipment-supplier basket. Pure Saudi exposure as a single-thesis allocation is more concentrated than most institutional portfolios should run, but Saudi-as-a-component of a broader AI infrastructure allocation is well-positioned for the rest of the decade.
Conclusion
The Saudi data center build-out is the most consequential AI infrastructure story in the Middle East and North Africa region for the rest of this decade. The capacity numbers — 250 megawatts in 2024, 400 to 500 megawatts in 2026, 1,300 megawatts targeted by 2030 — describe a structural reordering of regional cloud and AI infrastructure that has consequences well beyond the Kingdom’s borders. The capital underwriting the build-out — Humain’s 40-billion-dollar PIF commitment, ACWA Power’s 5-to-7-billion-dollar joint venture with Aramco, stc’s 4-billion-dollar Center3 expansion — is sovereign-scale and patient. The power-cost advantage at 3 to 5 cents per kilowatt-hour against 8-to-12-cent US and 15-to-20-cent Asian alternatives is structurally reproducible. The hyperscaler partnerships with Google, Microsoft, and Oracle anchor durable demand. The Arabic-language AI training infrastructure mandate creates a distinct workload pool that Saudi Arabia is positioned to capture.
For the foreign infrastructure investor, the right posture in 2026 is engaged and active. Position across ACWA Power, stc, and Mobily on the listed-equity side. Track the hyperscaler commitments through Alphabet, Microsoft, and Oracle. Monitor the chip-supplier names — NVIDIA, AMD — for Saudi-driven order momentum. Read the milestone calendar quarter-by-quarter and adjust position sizing as the 2026, 2027, and 2028 commissioning sequence delivers (or slips). The investors who put in the analytical work now will be the ones participating from the inside as the Saudi data center industry transitions from its early build-out phase into the mature operating cycle that will define the second half of this decade. The megawatts are real, the capital is committed, the demand is anchored, and the build is happening. The question is no longer whether Saudi Arabia will be a major data center hub by 2030 — it is how much of the value chain foreign capital captures along the way.
