Saudi Arabia’s AI Buildout Faces a Gap Between Announced and Live Power
Saudi Arabia’s AI infrastructure plans are expanding far faster than its operating data center load. The kingdom had 467 MW of data center capacity drawing power in the first quarter of 2026, compared with a planned pipeline of 3.4 GW and HUMAIN’s 6.6 GW target by 2034.
- Saudi Arabia had 467 MW of operational data center load in Q1 2026.
- The kingdom’s 2030 national target is 1.5 GW.
- HUMAIN has disclosed a 6.6 GW pipeline by 2034.
- Grid energization and anchor-tenant conversion are separate constraints.
- HUMAIN’s twin 100 MW campuses had no verified go-live by end-July 2026.
The gap between operating capacity and announced plans
According to figures disclosed by Saudi Arabia’s Ministry of Communications and Information Technology, the country operated 68 MW of data center capacity in 2021, 440 MW at the end of 2025 and 467 MW in the first quarter of 2026. The capacity was spread across more than 60 facilities.
Those figures sit well below the kingdom’s 1.5 GW national target for 2030. Planned capacity tracked at 3.4 GW in the second quarter of 2026, while HUMAIN, the sovereign AI company owned by the Public Investment Fund, has disclosed a pipeline of 6.6 GW by 2034 and an interim goal of 1.9 GW by 2030.
The source distinguishes between three measures that are often grouped together in coverage of the sector. Operational load is electricity actually contracted and drawn. National targets are policy commitments, while corporate pipelines represent land, capital and intent. The difference matters because an announced gigawatt does not become an operating megawatt until it is connected to firm power and supported by a tenant.
Grid connections are the first major constraint
The source describes Saudi Arabia’s main supply-side challenge as allocation and delivery rather than generation in aggregate. New AI facilities must pass through a feasibility study, capacity review, interconnection point, substation expansion and final grid acceptance.
Tenders ran in the first half of 2026 for a 2,000 MVA bulk supply point and two substations, described as physical preconditions for HUMAIN’s dense AI load. Transformer lead times were cited as two to four years across the global supply chain.
HUMAIN’s own sequence of agreements also points to power as an early requirement. The company signed an xAI framework covering 500 MW of power and $3 billion in equity in November 2025, followed by a joint venture with stc’s center3 covering up to 1 GW in December 2025. In January 2026, it signed a $12 billion financing framework with the National Infrastructure Fund. Chip and application agreements with AMD, Microsoft, Qualcomm and Adobe followed around the 2026 LEAP cycle.
Large anchor deals must still convert into live demand
The second constraint is demand concentration. Saudi Arabia’s AI expansion depends on a relatively small number of sovereign-scale commitments converting from announcements into contracts and energized load.
HUMAIN broke ground on twin 100 MW campuses in Riyadh and Dammam in August 2025, with a committed go-live in the second quarter of 2026. The source says that, as of the end of July 2026, the go-live of either campus could not be verified from a filing or credible report, and that the company had declined to disclose delivery progress when asked.
Other commitments also remained at different stages. The stc and center3 joint venture targeting up to 1 GW had been extended but was not signed as of a June 2026 filing. DataVolt’s Riyadh East project, planned at 44 MW with 36 MW anchored by HUMAIN, had not closed financing. AWS committed $5.3 billion to its Saudi region and an AI zone with HUMAIN carrying 50 MW by 2028.
What the figures mean for Saudi Arabia’s AI market
The two bottlenecks operate independently. Grid energization depends on procurement, equipment manufacturing and utility scheduling, while anchor conversion depends on counterparties’ capital cycles and global capacity plans. Strong deal activity therefore does not prove that power is available, just as new substations do not guarantee signed tenants.
The latest operating figures show the challenge. Capacity increased by 27 MW between the end of 2025 and the first quarter of 2026. The source calculates that reaching the 1.5 GW 2030 target would require roughly 260 MW of net additions per year from 2027 onward, assuming the target is pursued on that schedule.
For operators and investors, the relevant evidence is therefore shifting from headline gigawatts to substations, interconnection progress, financing closures, signed tenants and billed power. Sovereign AI demand can give HUMAIN priority access to major workloads, but the source says participants without anchor tenants may face later energization and a market in which the sovereign champion sets terms.
ACN Newswire: Artificial Intelligence: Gigawatts Are Announced. Megawatts Must Be Energized.

