Aiming to be a global leader in the field of artificial intelligence, the Kingdom of Saudi Arabia has initiated a huge national project to develop sovereign AI factories that run on the world-leading computing infrastructure. Humain, a new British-based public-sector AI firm founded by the Kingdom of Saudi Arabia Public Investment Fund (PIF), is at the center of this transformation. Announced in May 2025, Humain is charged with building and running the next generation of data centers in the country, training fundamental large language models, and hastening the evolution of sovereign AI systems available to both government and businesses.
It is not the intention of Saudi Arabia to merely embrace AI, but to also have control of its development pipelines, compute power, and technology stack. This initiative is made at the most opportune moment when AI leadership is becoming synonymous with geopolitical power, economic strength, and national safety. The difference is in the size, speed and sovereignty of the first design of Saudi Arabia's plans on AI infrastructure.
Source: Global News Wire
National Vision of Saudi Arabia
The campaign to have sovereign AI factories is based on the Vision 2030 of Saudi Arabia, which is a national economic agenda that seeks to make Saudi Arabia resilient by reducing dependence on oil by investing in digital infrastructure, advanced technology, and human capital. The key element in this approach is the Saudi Data and Artificial Intelligence Authority (SDAIA), which was created in 2019. SDAIA is also in charge of the national data management, governing machine learning policy and collaboration in line with its National Strategy on Data and Artificial Intelligence (NSDAI).
Humain works under this umbrella and its mission is to create sovereign computing systems that enable Saudi Arabia to own its AI infrastructure, store and process data locally and train language models specific to Arabic, Islamic, and Middle Eastern contexts. It has among its highest priorities to create a multimodal Arabic large language model (LLM) to drive applications in government, healthcare, education, and defense.
Such a sovereign-first approach will lessen reliance on external technology platforms and cloud companies so that the Kingdom will remain in control of essential AI systems and the data that educates them.
Mega Infrastructure
The core of the Saudi plan is the building of AI factories, which are huge hyperscale data centers operated to perform massive training and inference jobs with the help of advanced graphics processing units (GPUs). The initial stage will consist of an 18,000-chip supercomputer that will be based on the NVIDIA Grace Blackwell (GB200) GPUs and high-bandwidth InfiniBand networks. This cluster will produce multi-exaflop performance and will be the basis of training future AI in language modeling, generative media, simulation, and autonomous systems.
In the next five years, Humain will increase its infrastructure to 500 megawatts (MW) of AI compute capacity, a standard that few countries have ever aimed to reach in the past. A comparison of power can be made; 1 megawatt can operate hundreds of GPUs, and 500 MW is the compute required to support millions of machine learning applications concurrently. This will be capable of not only serving local projects in AI but also providing on-demand compute to outside collaborators in countries such as Germany or China.
The AI factories are designed using sustainable power as their intent, using solar, wind power and high-efficiency grid power to reduce carbon emissions. The data centers will also be linked through high-velocity fiber-optic cables to the larger telecom nodes, which will allow them to access the edge devices, research labs and commercial applications within the Kingdom at high-velocity.
Strategic Technology Deals with U.S. Tech Giants
The fast development of AI capacity in Saudi Arabia would be impossible without the strategic collaboration with the top semiconductor and cloud computing companies of the U.S. In May 2025, Saudi authorities inked more than $23 billion in agreements with NVIDIA, AMD, Amazon Web Services (AWS), and Qualcomm to provide chips, software platforms, and cloud services to run its sovereign AI factories.
NVIDIA is in the middle of such activities. The chip giant in cooperation with Humain, is supplying tens of thousands of its newest GB300 Blackwell GPUs, starting with an initial 18,000-unit introduction. Such GPUs are specifically made to train large AI models effectively and will act as the backbone of Saudi data centers. NVIDIA is also providing software such as Omniverse and NeMo to speed up simulation, modeling, and LLM fine-tuning.
Likewise, AMD joined a 10-billion-dollar alliance with Humain to deliver high-performance Instinct MI300 GPUs and EPYC server CPUs, and open-source software through the ROCm platform. This two-vendor approach is likely to enable Saudi Arabia to distinguish chip supply sources and take the most valuable cost-performance trade-offs in various AI tasks.
AWS is opening the first of its kind AI Zone in Saudi Arabia with an investment of over $5 billion to provide local AI cloud services. These are access to special GPU instances, low latency networking and sovereign data control to be in compliance with national regulations. AWS will also provide training and certification to the local talent under its cloud academy programs.
Qualcomm is also contributing in the form of a Memorandum of Understanding (MoU) to enable Saudi Arabia to create a national edge-AI ecosystem. A Qualcomm Design Center will be opening in Riyadh to design mobile and embedded AI chips. It will hire 500 engineers and help to port the Arabic LLMs of Saudi Arabia to run effectively on edge devices.
These alliances, with most being announced at a U.S.-Saudi investment summit in May 2025, symbolize an enormous shift in the world of artificial intelligence. The transactions make Saudi Arabia not only a consumer of AI, but also a developer of next-generation compute infrastructure.
Implications
The consequences of the Saudi sovereign AI strategy are huge. Economically, AI is expected to add more than 130 billion to Saudi GDP by 2030 or approximately 12 percent of its total economic output. Data center operations, AI research, software development and chip design are some of the areas that are likely to create thousands of jobs. It is not only attracting venture capital and startup talent, as Saudi Arabia ranked as the leading emerging market in Q1 2025 in terms of venture capital funding.
Geopolitically, the AI expansion in Saudi Arabia is a straightforward message: the nation is spending on long-term digital sovereignty. Hosting compute infrastructure in the country means that sensitive information is not exposed to foreign control. The move also puts Saudi Arabia in a position of strength in international AI negotiations, particularly as nations compete to put in place norms on AI safety, governance and trade.
It is a strategic timing. With the increase in the U.S. China tech tensions, numerous American companies are seeking to find amicable partners to diversify semiconductor supply chains. Saudi Arabia provides political support, financial resources and increasing regional power. Simultaneously, the Kingdom is structuring relations with AI companies across the world, such as OpenAI or xAI, as well as European and Southeast Asian startups.
Although the AI factories will mainly cater to national requirements, Saudi Arabia is also gearing up to be a net exporter of AI services. As computing prices increase in the U.S. and Europe because of power constraints and regulation, Saudi Arabia, with cheap energy and vast land, is an attractive place to host and train AI worldwide.
Looking Ahead
In the future, Humain intends to grow even bigger. It has a long-term plan to have 6.6 gigawatts (GW) of total AI compute by 2034, or approximately ten times its original target, which is estimated to cost it $77 billion in infrastructure investment. The objective is to be the home of 7 percent of the global AI workloads by 2030, which will make Saudi Arabia among the top five countries in the world in terms of AI compute capacity.
This vision is not however challenge free. The huge data centers need massive power, water and cooling systems. Despite the priority of sustainability, the environmental issues may accumulate, particularly in dry areas. The other bottleneck is talent. It will be important to maintain momentum by training and retaining thousands of engineers, data scientists and AI researchers.
The issue of data governance is sensitive as well. Open science and interoperability standards may be incompatible with sovereign control, particularly when the training data includes sensitive data or data that is not shared. Saudi Arabia needs to balance between sovereignty and world cooperation.
Another risk that is essential is cybersecurity. Cyberattacks will target AI factories as a strategic point. To ensure that national AI infrastructure is not being spied on or sabotaged, Humain and SDAIA will have to spend a lot of money on encryption, threat detection and resilience engineering.
Nevertheless, the way is clear despite these obstacles. Saudi Arabia does not want to be a consumer of AI. The Kingdom is developing infrastructure, talent, and policy environment to influence AI development from the ground up through Humain and a closely knit public-private model. Its sovereign AI factories can easily serve as an example to other countries that want to ensure their digital future.
.svg){kind=icon}
