Humain quadruples capacity in deal exceeding 689 percent of MIS revenue as Allam deployment accelerates toward a gigawatt of computing

Humain, the artificial intelligence unit owned by Saudi Arabia's Public Investment Fund, has awarded new contracts that will quadruple its data centre capacity. According to public disclosures, Tadawul-listed Al Moammar Information Systems (MIS) will construct 200 megawatts of data centre capacity, adding to ongoing work on a 50-megawatt project, bringing total built capacity to a quarter of a gigawatt.

Development will proceed across multiple engineering, procurement, and construction phases, with final contracts expected within two weeks. Although the total value was not specifically disclosed, MIS confirmed that the contracts exceed 689 percent of its 2025 revenue of 1.3 billion Saudi riyals, highlighting the scale of the investment and its direct financial impact on the technology infrastructure sector.

Moving from experimentation to building 200 megawatts of computing capacity reflects a structural shift in sovereign AI spending efficiency.The physical expansion coincides with a partnership between Humain and Microsoft to offer the Allam large language model to users and enterprise clients, pairing physical infrastructure growth with the global cloud distribution of Arabic language models.

The move comes alongside a report by consultancy Alvarez & Marsal estimating that AI and cloud expansion in Saudi Arabia could require up to 42 billion dollars in capital by 2030, including around 32 billion dollars in debt financing, if roughly half the announced data centre capacity is delivered. Under the report's baseline scenario, installed data centre capacity in the kingdom is projected to rise from 410 megawatts today to 1 gigawatt by the end of the decade.

This operational shift alters practical calculations for regional institutions and tech firms. Building hyperscale domestic data centres inside Saudi Arabia lowers latency for sovereign Arabic models such as Allam and reduces reliance on rented cross-border computing. For banks, government agencies, and enterprises across the Gulf, Egypt, and the Levant subject to strict data sovereignty and local processing rules, bringing advanced compute online makes it feasible to build large-model applications with more predictable operational and inference costs, encouraging technical teams to migrate enterprise workloads to locally hosted cloud environments.

Reaching the gigawatt threshold requires regional technology leaders to align their software architectures with local inference platforms rather than depending on distant servers.The accelerated delivery of large-scale computing projects creates a practical demand for advanced skills in data centre operations, power optimization, and the integration of sovereign models into commercial products, making abundant compute a decisive factor in driving automation across the Middle East.