Kholoud Hussein
Saudi Arabia’s artificial intelligence ambitions are entering a new phase, with HUMAIN moving beyond its initial role as a state-backed AI infrastructure developer to build a broader business spanning data centers, computing capacity, Arabic-language models, enterprise software, strategic investments, and physical AI applications.
Since its launch by the Public Investment Fund (PIF) in May 2025, HUMAIN has assembled partnerships with some of the world’s largest technology companies, including NVIDIA, AMD, Cisco, Qualcomm, Microsoft, Amazon Web Services (AWS), and Mistral AI. It has also expanded its relationship with Elon Musk’s xAI, invested $3 billion in the company, and announced plans to develop large-scale computing infrastructure in Saudi Arabia.
The scale of its ambitions has increased sharply. In an interview published in October 2026, HUMAIN CEO Tareq Amin said the company was targeting 1.7 gigawatts (GW) of data center capacity in 2027, around 3 GW by 2030, and approximately 6 GW by 2034. The targets place HUMAIN at the center of Saudi Arabia’s effort to become a major global destination for AI computing.
Yet the company’s long-term value will depend on more than the amount of computing power it can install. HUMAIN must also secure financing, convert infrastructure into contracted demand, develop commercially competitive AI products, attract technical talent, and help Saudi businesses adopt AI at scale. Its strategy reflects a broader economic question facing the Kingdom: how to turn substantial investment in technology infrastructure into a durable source of productivity, intellectual property, and non-oil growth.
From national AI platform to global technology business
HUMAIN was established to operate across the AI value chain, rather than concentrate on a single product or service. Its mandate encompasses next-generation data centers, high-performance computing and cloud platforms, advanced AI models, and AI-powered applications for businesses and government entities.
The company’s creation reflects the PIF’s wider strategy of investing in industries that can diversify the Saudi economy, attract international capital, and build domestic technological capabilities. It also brings AI infrastructure and software development into a coordinated commercial structure, linking the Kingdom’s energy resources and investment capacity with global semiconductor suppliers, cloud providers, and model developers.
This approach addresses a structural challenge in the AI market. Building advanced models requires substantial computing resources, specialized chips, reliable power, and high-speed networking. Operating them commercially also requires cloud platforms, software, enterprise integration, and customers willing to pay for the resulting services.
HUMAIN is attempting to connect these layers within one business. Its partnerships provide access to technology and expertise that would be difficult to reproduce independently in the short term, while its Saudi operations are intended to establish a local base for computing, model development, and commercial deployment.
The strategy is not limited to serving Saudi government agencies. HUMAIN has repeatedly positioned the Kingdom as a potential exporter of AI computing and services, serving regional and international customers from infrastructure located in Saudi Arabia. The opportunity rests partly on the country’s access to energy, land, capital, and connectivity, as well as its ambition to develop locally relevant AI products.
The latest deals: Infrastructure, models, and real-world applications
HUMAIN’s recent announcements show how its strategy is expanding from building computing capacity toward attracting customers, developing products, and investing in technologies with global markets.
One of the most significant developments came on August 31, 2026, when HUMAIN, AMD, and Cisco announced that AI computing infrastructure powered by AMD Instinct MI355X GPUs, AMD EPYC CPUs, and Cisco Silicon One networking was live in Saudi Arabia and serving customers.
The companies also outlined plans to deploy up to 250 megawatts (MW) of additional AI infrastructure in the next phase, beginning in 2027. Their joint venture remains on track to deploy up to 1 GW of AI infrastructure by 2030, according to the announcement.
The commercial significance is that HUMAIN is beginning to demonstrate operational computing services rather than relying solely on future capacity announcements. Production workloads can help establish whether its infrastructure meets customer requirements for performance, availability, and cost. They can also generate demand data that informs subsequent investment decisions.
Another major development involves DataVolt, a Saudi-based developer and operator of sustainable digital infrastructure. On August 31, the companies announced an expanded partnership to develop 100 MW of an AI-ready data center project at Oxagon, NEOM’s industrial city, as part of a 360 MW first phase of DataVolt’s planned campus. Construction was reported to be underway.
The project illustrates the importance of partnerships in delivering AI infrastructure. Data center development requires land, power connections, cooling systems, construction expertise, and specialized equipment, in addition to computing hardware. Sharing responsibilities across infrastructure and technology partners can help distribute the cost and execution risks of a capital-intensive buildout.
HUMAIN is also pursuing partnerships that connect infrastructure with the software layer. In August 2026, Microsoft and HUMAIN announced the first milestone in a long-term strategic collaboration focused on bringing HUMAIN’s ALLAM models into Microsoft’s AI ecosystem, including Microsoft Foundry and the Microsoft 365 Copilot ecosystem. The collaboration also envisages joint work between HUMAIN specialists and Microsoft’s Forward Deployed Engineers to identify and deploy AI solutions for organizations.
The arrangement could help HUMAIN reach enterprise customers through an established global software ecosystem. It also highlights a commercial challenge: developing an AI model is only one part of building a sustainable business. Models must be integrated into workflows, connected to organizational data, secured, and adapted to the needs of individual customers.
A further partnership with French AI developer Mistral AI, announced on August 24, 2026, focuses on infrastructure, advanced model development, and AI deployment in Saudi Arabia and the wider region. Initial areas of interest include cybersecurity and voice applications, alongside the development and localization of advanced models with strong Arabic-language performance. The companies described the collaboration as involving hundreds of millions of euros and said Mistral would explore using HUMAIN’s data center infrastructure.
These partnerships serve different purposes. AMD and Cisco support computing infrastructure; DataVolt supports data center development; Microsoft provides a route into enterprise software; and Mistral brings model-development capabilities. Together, they suggest that HUMAIN is seeking to build a connected commercial ecosystem rather than a collection of isolated technology projects.
The $3 billion xAI investment and HUMAIN’s global ambitions
HUMAIN’s investment strategy extends beyond infrastructure partnerships. In February 2026, the company announced a $3 billion investment in xAI’s Series E financing round, becoming a significant minority shareholder. Following xAI’s acquisition by SpaceX, HUMAIN’s holdings were converted into SpaceX shares, according to the company’s announcement.
The investment builds on an earlier agreement between HUMAIN and xAI to develop more than 500 MW of AI data center capacity in Saudi Arabia and deploy Grok across the Kingdom. The partnership links access to advanced AI models with the development of domestic computing infrastructure.
The $3 billion investment is particularly notable because it demonstrates that HUMAIN is prepared to deploy capital into international AI platforms, not just Saudi-based assets. It could provide financial exposure to the growth of a major technology business while strengthening the relationship between HUMAIN and a partner involved in advanced AI models and large-scale computing.
However, the transaction should be distinguished from direct investment in Saudi infrastructure. An equity investment in a global technology company does not automatically create domestic computing capacity, jobs, or technology transfer. Those benefits depend on the associated commercial arrangements, local operations, and the extent to which the partnership creates opportunities for Saudi engineers and businesses.
HUMAIN has also been expanding its investment plans. In September 2026, Semafor reported that Amin intended to launch a large global venture capital fund, potentially larger than a previously contemplated $10 billion vehicle. According to the report, the fund would have offices in Saudi Arabia, the United States, France, and the United Kingdom. Amin told the publication: “We don’t do passive investments; that will never happen.”
The reported fund would focus on investments that contribute to HUMAIN’s strategic objectives, including companies willing to use Saudi data centers for some of their computing operations or bring staff to the Kingdom. If implemented on the scale reported, this could give HUMAIN a role in shaping where AI companies locate their computing workloads, technical teams, and commercial activities.
The fund remains a reported plan rather than a confirmed, fully capitalized investment vehicle. Its eventual size, investment mandate, and launch timetable will be important indicators of whether HUMAIN intends to become a major global AI investor alongside its infrastructure and software businesses.
Data centers: The backbone of HUMAIN’s expansion
Data centers sit at the center of HUMAIN’s business model because advanced AI requires large amounts of computing power. Training sophisticated models, operating generative AI applications, and serving large numbers of users all depend on access to specialized processors and the infrastructure needed to operate them reliably.
Saudi Arabia has been expanding its broader digital infrastructure base. According to the Saudi Press Agency, operational data center capacity rose from 68 MW in 2021 to more than 440 MW in 2025, while the Kingdom hosted more than 60 data centers developed by over 20 companies, with investments exceeding SAR16 billion.
HUMAIN’s targets would represent a substantial expansion beyond that existing market. Its October 2026 target of 1.7 GW in 2027, followed by approximately 3 GW by 2030 and 6 GW by 2034, would require a combination of new facilities, computing hardware, power infrastructure, cooling systems, networking, and long-term customer demand.
The cost implications are considerable. EnterpriseAM reported that Amin estimated AI data center capacity typically costs between $10 million and $12 million per MW. Applying that benchmark to 1.45 GW of additional capacity—the difference between 250 MW and a 1.7 GW target—would imply roughly $14.5 billion to $17.4 billion in infrastructure costs. This is an illustrative calculation, not a company-announced budget; actual costs would depend on project scope, equipment, financing, and what is included in the capacity target.
The company has already taken steps to establish financing channels. In January 2026, HUMAIN and Saudi Arabia’s National Infrastructure Fund (Infra) announced a strategic financing framework of up to $1.2 billion to support AI and digital infrastructure projects. The non-binding framework covers the development of up to 250 MW of hyperscale AI data center capacity. The parties also agreed to explore an AI data center investment platform designed to facilitate participation by domestic and international institutional investors.
Separately, Bloomberg reported in September 2026 that HUMAIN was exploring raising an initial $2.5 billion from global and domestic investors for a data center fund. The reported fund would support a 250 MW project being developed with Al Moammar Information Systems and could eventually expand its capacity. This remains a reported financing plan, distinct from the Infra framework.
The financing challenge is therefore not simply how much capital HUMAIN can attract, but how it structures that capital across projects with different risk profiles. Equity, debt, institutional investment, joint ventures, and long-term customer contracts can all play a role. The critical factor will be whether projected demand and operating revenue justify the cost of building capacity at the pace the company has outlined.
The energy advantage—and the infrastructure constraints
Saudi Arabia’s energy resources are central to its proposition as a global AI computing destination. Data centers require reliable electricity around the clock, and AI workloads can create particularly high power and cooling requirements. The availability and cost of electricity can therefore influence the competitiveness of a data center as directly as its computing hardware.
In an interview published in October 2026, Amin described HUMAIN’s network of AI campuses as spanning approximately 65 square kilometers and backed by 14 GW of grid capacity. He argued that the Kingdom’s combination of energy, land, connectivity, and the ability to deploy infrastructure at scale gives it a competitive advantage.
But available power capacity is not the same as operational data center capacity. Facilities must secure grid connections, cooling, networking, permits, equipment, and construction schedules. They must also be designed to maintain service reliability as workloads and hardware requirements evolve.
The economic opportunity extends beyond hosting servers. Building a domestic AI infrastructure industry can create demand for engineering, construction, cooling, power management, cybersecurity, and equipment maintenance. The PIF’s HUMAIN portfolio page also identifies local manufacturing opportunities in data center cooling equipment, power equipment, and server racks.
Localizing these supply chains could help retain more of the economic value generated by infrastructure investment. Without that development, a large share of the expenditure could flow to international hardware and technology suppliers, leaving Saudi Arabia primarily as the location where the equipment operates.
HUMAIN will therefore need to balance rapid deployment with efforts to develop domestic suppliers, technical capabilities, and more efficient operations. Energy availability is a competitive advantage, but converting it into a lasting industrial advantage requires more than access to electricity.
Closing gaps in Arabic AI, sovereign computing, and enterprise adoption
HUMAIN’s ambitions are not limited to data centers. Its model strategy addresses another gap: the need for advanced AI systems that perform well in Arabic and reflect the language’s linguistic, cultural, and commercial context.
Arabic presents challenges for AI developers because of its regional dialects, morphological complexity, and the differences between formal written Arabic and everyday speech. Organizations also need models that can work with sector-specific terminology and sensitive data while meeting local requirements for security, privacy, and governance.
HUMAIN’s ALLAM models, developed in collaboration with the Saudi Data and AI Authority (SDAIA), are intended to address Arabic-language needs. In August 2026, Microsoft and HUMAIN announced plans to bring ALLAM into Microsoft’s AI ecosystem, potentially giving organizations a more direct route to use the models in enterprise applications.
In September 2026, HUMAIN also unveiled humain-m3, an Arabic-language model commissioned by HUMAIN and developed by MiniMax. According to the announcement, the model has 428 billion parameters, uses a mixture-of-experts architecture, and was further pre-trained on more than one trillion tokens of Arabic-native content. It entered research preview through HUMAIN Node, the company’s platform for developers, researchers, and enterprises seeking access to AI models and inference capabilities. HUMAIN reported that the model achieved the highest average performance among the frontier models evaluated across seven public Arabic benchmarks. That performance claim is based on the company’s announcement and should be assessed alongside independent evaluations and real-world deployments.
This illustrates HUMAIN’s hybrid approach: developing local AI capabilities while working with international model developers. The objective is not necessarily to build every component from scratch, but to use partnerships to accelerate development while retaining the ability to customize, deploy, and commercialize AI for Saudi and regional needs.
Another gap concerns the transition from experimentation to routine business use. Companies may have access to AI tools but lack the technical expertise, secure data environments, or internal processes required to deploy them effectively. HUMAIN’s collaboration with Microsoft is designed partly to address this challenge through direct engineering support for organizations identifying and implementing AI use cases.
Its enterprise operating system, HUMAIN ONE, also reflects a shift toward AI agents: software systems designed to carry out tasks and coordinate workflows rather than simply generate text or answer questions. In May 2026, HUMAIN announced an expanded collaboration with AWS around HUMAIN ONE, aiming to support organizations in building, deploying, and governing autonomous AI agents at scale.
The commercial test will be whether these products deliver measurable improvements in cost, productivity, service quality, and decision-making. A strong model or platform is not enough on its own; it must solve a problem that customers consider important enough to fund.
Beyond the data center: Physical AI and industrial applications
HUMAIN’s next stage could also involve extending AI from digital services into physical operations. This is important for an economy with substantial activity in energy, logistics, manufacturing, mining, and large infrastructure projects.
On August 31, 2026, HUMAIN and Applied Intuition announced a strategic collaboration to deploy physical AI in Saudi Arabia, beginning with autonomous trucking. The companies said they planned to build a national-scale autonomous logistics network, with thousands of autonomous trucks operating across key Saudi freight corridors by 2030. The partnership could later extend into robotaxis, ports, mining, manufacturing, agriculture, and construction.
The collaboration links HUMAIN’s AI infrastructure with Applied Intuition’s autonomy technology. Its potential economic impact lies in applications such as increasing fleet utilization, improving freight capacity, reducing operating costs, and supporting logistics networks that can operate more efficiently.
The timeline and scale are ambitious, and the announced target should not be mistaken for an already operational national fleet. Deployment will depend on technical performance, road and logistics infrastructure, safety standards, regulation, and commercial economics.
Still, the move signals that HUMAIN is seeking opportunities where AI can influence physical productivity, not only digital services. If such applications prove commercially viable, they could help connect Saudi AI investment to improvements in existing industries rather than treating AI as a separate technology sector.
What comes after Vision 2030?
HUMAIN’s immediate targets are tied to the period leading up to 2030, but its 6 GW capacity ambition by 2034 points to a strategy extending beyond the Vision 2030 horizon. The company’s longer-term objective appears to be building a Saudi-based AI industry capable of serving customers globally, supported by infrastructure, models, software, and investment.
For that ambition to become economically sustainable, several priorities will matter.
First, infrastructure must be matched with demand. Data centers are capital-intensive assets, and unused computing capacity can weaken returns. Long-term contracts, diversified customers, and strong utilization rates will be essential to support the expansion.
Second, HUMAIN must turn partnerships into capabilities that can be developed and commercialized locally. International technology providers can accelerate access to chips, models, and software, but the long-term objective should include Saudi expertise in system design, AI engineering, product development, and intellectual property.
Third, the company will need to demonstrate that its AI products can compete on performance, reliability, security, and cost. Arabic-language models and locally hosted infrastructure may provide differentiation, but their value will ultimately depend on the quality of the products and the problems they solve.
Fourth, financing needs to remain aligned with commercial returns. The $1.2 billion Infra framework, the reported $2.5 billion data center fund, and plans for a much larger global venture capital vehicle represent different funding channels and ambitions. They should not be treated as one consolidated investment commitment. HUMAIN has not publicly established a single, fully funded budget covering all its announced capacity targets and investment plans.
Finally, Saudi Arabia will need to deepen its talent base and domestic supply chain. Data centers can be built with international equipment, and models can be developed with global partners, but a competitive AI economy requires engineers, researchers, entrepreneurs, and product teams able to create and improve technology over time.
In a recent interview, Amin described the broader ambition in terms of building an industry from Saudi Arabia rather than merely bringing AI into the country. “Our ambition is not simply to bring AI into Saudi Arabia. It is to build an AI industry from Saudi Arabia that can compete globally,” he said.
That distinction captures the economic stakes. If HUMAIN succeeds, Saudi Arabia could become more than a market for imported AI products or a location for international data centers. It could develop into a producer and exporter of computing services, Arabic AI models, enterprise applications, and industrial AI solutions.
If the company’s expansion is driven mainly by capacity targets without equivalent progress in customer demand, local capabilities, and commercially successful products, the economic returns could fall short of the investment scale.
The next phase will therefore be measured not only in megawatts and gigawatts, but also in contracted workloads, paying customers, productivity gains, locally developed intellectual property, and technology exports. Those outcomes will determine whether HUMAIN becomes a lasting pillar of Saudi Arabia’s post-2030 economy—or primarily an ambitious infrastructure platform in a rapidly changing global AI market.
