Gulf States AI: The $100 Billion Desert Bet on Inhuman Intelligence

The Gulf states are making the largest coordinated bet on artificial intelligence by any group of nations in history. Saudi Arabia, the United Arab Emirates, Qatar, and their smaller neighbors have collectively committed over $100 billion to AI infrastructure, partnerships, research institutions, and venture investments. They are attempting to purchase, in a single generation, the AI capability that the United States and China have built over decades of academic research, corporate innovation, and gradual institutional development.

This is either visionary statecraft or the most expensive case of technology FOMO ever recorded. The answer depends on whether sovereign AI capability can be bought, or whether it must be grown organically through the slow accumulation of talent, institutional knowledge, and research culture. The Gulf states are running the experiment at unprecedented scale, and the results will reshape the global AI landscape regardless of the outcome.

INHUMAIN.AI exists, in part, because of this experiment. Our name is a direct response to HUMAIN, Saudi Arabia’s national AI company, which appropriated the word “humain” — a term that should belong to humanity, not to a sovereign wealth fund. We track every HUMAIN deal, partnership, and milestone at HUMAIN Watch. But the Gulf AI story is far larger than a single company. This analysis covers the full landscape.

---

Saudi Arabia: The Kingdom’s AI Pivot

Saudi Arabia’s AI strategy is inseparable from Vision 2030, Crown Prince Mohammed bin Salman’s plan to diversify the Saudi economy away from petroleum dependence. AI is positioned not merely as one sector among many but as the foundation of the entire diversification effort — the technology that will power the smart cities, the autonomous logistics networks, the digital government services, and the knowledge economy that Vision 2030 envisions.

The institutional architecture is multi-layered:

HUMAIN (National AI Company)

Dimension	Detail
Founded	May 2025 (announced at PIF Private Sector Forum)
Owner	Public Investment Fund (PIF)
CEO	Tareq Amin (ex-Rakuten Mobile CTO)
Chairman	Crown Prince Mohammed bin Salman
Regulator	SDAIA (Saudi Data & AI Authority)
Stated Mission	Build Saudi Arabia into a global AI power
Original Domain	humain.ai (now humain.com)

HUMAIN is the operational vehicle for Saudi AI ambitions. It functions as a holding company, infrastructure builder, and strategic investor simultaneously. Its partnership portfolio is extensive and growing:

Partner	Deal Type	Estimated Value	Date
NVIDIA	GPU supply + sovereign AI infrastructure	Multi-billion	May 2025
AMD + Cisco	1GW AI infrastructure joint venture	Multi-billion	Nov 2025
xAI (Elon Musk)	$3B equity investment	$3B	2025
Groq	Inference infrastructure	$1.5B commitment	Dec 2025
Amazon	Strategic partnership	Undisclosed	2025
Qualcomm	Edge AI partnership	Undisclosed	2025
Scale AI	Data labeling and AI infrastructure	Undisclosed	2025
Cerebras	Wafer-scale AI compute	Undisclosed	2025

The scale of investment is staggering. HUMAIN’s data center buildout alone represents a commitment of tens of billions of dollars. The company is building or planning at least 11 facilities across Saudi Arabia, with a target compute capacity that would rank among the largest in the world.

For detailed, continuously updated tracking, see: HUMAIN Watch.

SDAIA (Saudi Data and AI Authority)

SDAIA is the regulatory and policy body governing AI in Saudi Arabia. Established in 2019, it reports directly to the Council of Ministers and is chaired by Crown Prince Mohammed bin Salman. SDAIA published the National Data Management Office’s Personal Data Protection Law (PDPL), which came into effect in September 2023, and oversees AI ethics and governance frameworks.

SDAIA also operates the National Data Bank and has been instrumental in establishing data sharing agreements between government entities — creating the kinds of unified datasets that AI systems require for training on Saudi-specific applications.

KAUST and Academic Institutions

King Abdullah University of Science and Technology (KAUST) is Saudi Arabia’s primary AI research institution, with significant investment in machine learning, computer vision, and natural language processing. KAUST has attracted international faculty and maintains research partnerships with leading global institutions.

However, KAUST’s research output, while respectable, does not approach the frontier. Saudi Arabia produces fewer top-tier AI research publications than individual departments at Stanford, MIT, or Tsinghua. The gap between infrastructure spending and research capability is one of the central tensions in the Saudi AI strategy.

NEOM

NEOM, the $500 billion megacity project in northwestern Saudi Arabia, is positioned as a deployment testbed for AI at city scale. The project envisions AI-managed transportation, energy systems, healthcare, and urban services. If built as planned, NEOM would be the most AI-saturated urban environment on Earth.

But NEOM has faced persistent criticism, delays, and skepticism. The project’s scale has been reduced from its original ambitious projections. Forced relocations of the Howeitat tribe from the project site resulted in at least one death and multiple prison sentences, drawing international condemnation. NEOM’s relationship to AI is currently more aspirational than operational.

---

United Arab Emirates: The Diversified Play

The UAE has pursued a more diversified and arguably more sophisticated AI strategy than Saudi Arabia, building multiple institutional pillars rather than concentrating investment in a single national champion.

G42

G42 (Group 42) is the UAE’s most prominent AI company, headquartered in Abu Dhabi and chaired by Sheikh Tahnoon bin Zayed Al Nahyan — the UAE’s national security advisor, brother of President Mohamed bin Zayed, and one of the most powerful figures in the Gulf.

G42 has navigated the US-China technology rivalry more visibly than any other Gulf entity. The company originally maintained significant partnerships with Chinese technology firms, including Huawei and BGI Group (the genomics company). Under US pressure, G42 divested its Chinese partnerships in 2024 and signed a strategic partnership with Microsoft, positioning itself firmly in the American technology orbit.

G42 Milestone	Detail
Microsoft Partnership	Strategic investment and technology partnership (2024)
Chinese Divestment	Exited partnerships with Huawei, BGI under US pressure
Cerebras Partnership	AI compute infrastructure collaboration
Core42	Cloud and infrastructure subsidiary
Inception	AI venture capital subsidiary
Revenue	Not publicly disclosed; estimated in billions

G42’s pivot from Chinese to American technology partnerships illustrates a fundamental reality of Gulf AI strategy: the Gulf states can build infrastructure and write checks, but they depend on technology partners for the actual AI capability. And the choice of technology partner is ultimately a geopolitical alignment decision, not merely a commercial one.

Technology Innovation Institute (TII)

TII, based in Abu Dhabi and part of the Advanced Technology Research Council, is the UAE’s primary AI research institution. Its most notable achievement is Jais, one of the first Arabic-language large language models, released in 2023.

Jais was a symbolically important milestone — the first competitive LLM designed for Arabic as a primary language rather than an afterthought. But it also highlighted the challenges of non-English AI development. Arabic is morphologically complex, has numerous dialects that diverge significantly from Modern Standard Arabic, and has a relatively smaller digital text corpus compared to English or Chinese. Jais required significant investment in data curation and linguistic expertise beyond what English-language model development demands.

TII has also worked on Falcon, a series of open-source language models that achieved strong benchmark performance upon release and demonstrated that a well-funded Gulf institution could produce competitive models. However, the Falcon models have not maintained pace with the rapid advancement of frontier labs in the US and China.

MBZUAI (Mohamed bin Zayed University of Artificial Intelligence)

MBZUAI, established in 2019, is the world’s first graduate-level university dedicated entirely to artificial intelligence. Located in Masdar City, Abu Dhabi, it offers Master’s and PhD programs in machine learning, computer vision, and natural language processing.

MBZUAI has recruited faculty from leading international institutions, including former researchers from Carnegie Mellon, MIT, and Oxford. The university provides full scholarships, stipends, and computing resources to attract talent from around the world.

The strategic logic is clear: if the Gulf states cannot organically produce AI talent at the required scale, they will build institutions to attract and train it. Whether MBZUAI can become a genuinely world-class research institution, or remains a well-funded but peripheral player, will take a decade or more to determine.

---

Qatar: The Smaller but Smarter Play?

Qatar’s AI investments are smaller than Saudi Arabia’s or the UAE’s, but arguably more focused. The Qatar Computing Research Institute (QCRI), part of Hamad Bin Khalifa University, has produced genuinely notable research in Arabic NLP, social media analysis, and disinformation detection.

Qatar’s strategy emphasizes specific domains rather than attempting to build general-purpose AI capability:

Domain	Initiative	Status
Arabic NLP	QCRI research programs	Active, internationally recognized
Smart infrastructure	Qatar Smart Program (Tasmu)	Connected to 2030 National Vision
Sports analytics	AI applications for FIFA/sports hosting	Active
Healthcare AI	Sidra Medicine partnerships	Active
Education	Qatar Foundation AI integration	Active

Qatar’s approach may be more realistic than Saudi Arabia’s or the UAE’s attempts at comprehensive AI sovereignty. By focusing on specific domains where it has natural advantages — Arabic language, regional infrastructure, niche research — Qatar avoids the trap of trying to compete with the US and China across the full AI capability spectrum.

---

Bahrain, Kuwait, and Oman: The Supporting Cast

The smaller Gulf states have more modest AI ambitions but are not absent from the landscape.

Bahrain has positioned itself as a regulatory sandbox and fintech hub, with AI applications in financial services as a priority. The Bahrain Economic Development Board has actively courted AI startups and established partnerships with Amazon Web Services and other cloud providers.

Kuwait has announced various AI strategy documents but has invested relatively little compared to its larger neighbors. Kuwait Investment Authority, one of the world’s oldest sovereign wealth funds, has made some AI-related portfolio investments but has not pursued a HUMAIN-style operational AI strategy.

Oman has articulated AI ambitions as part of its Vision 2040 plan but faces more severe fiscal constraints than its oil-richer neighbors. Oman’s AI investments are focused on government digitization and logistics optimization for its port and free zone infrastructure.

---

Comparative Analysis: Gulf AI Strategies

Dimension	Saudi Arabia	UAE	Qatar	Bahrain
Primary vehicle	HUMAIN (national company)	G42, TII, MBZUAI (diversified)	QCRI (research-focused)	EDB (regulatory sandbox)
Investment scale	$50B+ (estimated)	$30B+ (estimated)	$5B+ (estimated)	<$1B
Approach	Buy-and-build	Diversified institutional	Domain-specific	Regulatory niche
Key tech partner	NVIDIA, AMD, xAI	Microsoft, Cerebras	Various	AWS
China exposure	Limited (post-US pressure)	Reduced (G42 divestment)	Moderate	Low
Indigenous research	Weak	Moderate (TII, Falcon, Jais)	Moderate (QCRI)	Weak
Talent strategy	Import via HUMAIN	MBZUAI + import	QCRI recruitment	Limited
Human rights risk	High	High	Moderate	Moderate
Energy advantage	Strong (cheap power)	Strong	Moderate	Moderate
Regulatory maturity	Developing (PDPL)	Developing	Developing	Most developed (sandbox)

---

The Energy Equation: Advantage and Curse

The Gulf states’ most frequently cited advantage in AI infrastructure is energy. AI training and inference consume enormous amounts of electricity, and the Gulf states have among the lowest energy costs in the world, thanks to subsidized natural gas and growing solar capacity.

But the energy advantage is more complicated than it appears:

Cooling costs. AI data centers generate enormous heat and require massive cooling systems. In the Gulf, where ambient temperatures regularly exceed 45 degrees Celsius (113 degrees Fahrenheit), cooling costs can consume 30-40% of a data center’s total energy budget, compared to 10-15% in temperate climates. This partially negates the advantage of cheap electricity.

Water consumption. Conventional data center cooling requires significant water, a resource the Gulf states are desperately short of. The region depends heavily on energy-intensive desalination for fresh water. Using desalinated water for data center cooling creates a water-energy feedback loop that increases costs.

Grid capacity. The scale of AI data center buildout envisioned by Saudi Arabia and the UAE will require significant expansion of electrical grid capacity. Current grid infrastructure was designed for residential and industrial loads, not for multi-gigawatt data center clusters. Grid expansion requires years of construction and billions in additional investment.

Renewable integration. The Gulf states have abundant solar resources, and Saudi Arabia’s NEOM project includes plans for large-scale solar and green hydrogen production. But AI data centers require consistent, 24/7 power, and solar is intermittent. Integrating AI infrastructure with renewable energy sources requires storage solutions that add cost and complexity.

The net energy equation is favorable for the Gulf states but not as overwhelmingly advantageous as their promotional materials suggest. The total cost of operating a large AI data center in Saudi Arabia may be 20-30% lower than in the United States, not the order-of-magnitude advantage sometimes implied.

---

The Talent Gap: The Achilles Heel

The Gulf states’ most critical vulnerability is talent. Building data centers and purchasing GPUs are capital allocation problems, and the Gulf states excel at capital allocation. But operating those data centers, training models on them, and conducting the research that advances the state of the art requires human expertise that cannot be purchased with a wire transfer.

The Gulf states have thin domestic AI talent pools. Saudi Arabia, the UAE, and Qatar combined have populations smaller than many individual Chinese provinces. Their university systems, while improving, do not produce AI researchers at the scale or quality of American, Chinese, or even Indian institutions.

The result is extreme dependence on imported talent. HUMAIN’s CEO, Tareq Amin, is a naturalized American citizen born in Iraq with his career primarily in Japan and the United States. The research teams at TII, MBZUAI, and QCRI are predominantly composed of expatriates from South Asia, the Middle East, and Western countries.

This creates several risks:

Retention. Expatriate workers in the Gulf are there primarily for compensation. They do not have citizenship rights, face cultural and social restrictions, and are vulnerable to the kafala sponsorship system (though Saudi Arabia and the UAE have reformed aspects of this system). When compensation advantages diminish or personal circumstances change, expatriate talent leaves. Building institutional knowledge and research culture requires decades of continuity.

Security. AI systems for government applications — defense, surveillance, critical infrastructure — require personnel with security clearances. Granting security clearances to foreign nationals creates obvious risks, but relying on a thin domestic talent pool for all security-sensitive AI work limits capability.

Cultural barriers. The Gulf states’ social restrictions — limitations on alcohol, mixed-gender socializing, religious expression, and LGBTQ+ rights — make them less attractive to the global talent pool that has other options. The most in-demand AI researchers can work virtually anywhere in the world. They generally choose locations that offer not only high compensation but also personal freedom and cultural amenity.

---

Human Rights: The Elephant in the Server Room

Any analysis of Gulf AI that omits human rights is incomplete and dishonest. The Gulf states’ human rights records are directly relevant to their AI strategies in at least three ways.

Surveillance applications. AI-powered surveillance systems have enormous potential for authoritarian social control, and the Gulf states have documented histories of deploying surveillance technology against dissidents, journalists, and activists. Saudi Arabia’s use of NSO Group’s Pegasus spyware against journalist Jamal Khashoggi’s associates was extensively documented. The UAE’s Project Raven, which employed former NSA operatives to conduct surveillance operations including against American citizens, was revealed by Reuters in 2019. Building sovereign AI infrastructure gives these governments the capability to conduct surveillance at a scale and sophistication that purchased tools like Pegasus cannot match.

Labor conditions. The physical infrastructure of Gulf AI — the data centers, the fiber optic networks, the power plants — is built by migrant workers who operate under labor conditions that international human rights organizations have extensively documented as exploitative. The kafala system, wage theft, dangerous working conditions, and restrictions on workers’ freedom of movement remain concerns despite reform efforts. The contrast between the gleaming futurism of HUMAIN’s promotional materials and the conditions of the workers constructing its data centers is a tension that the Gulf states have not resolved.

Content moderation and censorship. AI systems deployed in the Gulf states will operate within the constraints of local censorship laws that restrict political speech, religious expression, and reporting on government activities. Language models trained or fine-tuned for Gulf deployment will embed these restrictions in their behavior. The question of whether Western technology partners are complicit in building AI systems designed to enforce censorship is a question that Microsoft, NVIDIA, and others have not adequately answered.

These concerns are not peripheral. They go to the core question of what AI is being built for. If the Gulf states are building AI infrastructure to diversify their economies, improve government services, and advance scientific research, that is one story. If they are building AI infrastructure to perfect authoritarian governance at AI-enabled scale, that is a very different story. The truth is probably both, and the inability to separate the two is precisely the problem.

For our analysis of how HUMAIN specifically navigates these tensions, see: HUMAIN vs. INHUMAIN: Why We Exist.

---

Western Technology Partners: Complicity or Commerce?

Every major Western technology company is involved in Gulf AI. NVIDIA sells GPUs. Microsoft has invested in G42. Amazon has a strategic partnership with HUMAIN. AMD and Cisco are building joint ventures. Qualcomm, Groq, Cerebras, Scale AI — the list continues.

These partnerships raise questions that the companies involved prefer not to answer:

Due diligence. What safeguards exist to prevent AI infrastructure built with Western technology from being used for surveillance, military targeting, or human rights abuses? The companies involved generally cite contractual restrictions on end-use. But contractual restrictions are enforced through monitoring, and Western companies have limited visibility into how their technology is used once it is deployed in sovereign Gulf infrastructure.

Export control compliance. The US has expressed concern about advanced AI hardware being reexported from Gulf states to China or other restricted destinations. The October 2023 export control updates included provisions targeting Gulf reexport channels. Western companies operating in the Gulf must navigate complex compliance requirements while maintaining commercial relationships worth billions of dollars.

Reputation. As AI becomes more politically salient, the association between Western technology brands and authoritarian governments becomes more reputationally costly. Google’s experience with Project Dragonfly (the censored Chinese search engine that was abandoned after employee protests) suggests that there are limits to what technology workers and public opinion will tolerate. Whether Gulf AI partnerships approach those limits remains to be seen.

---

The $100 Billion Question

The Gulf states’ AI bet raises a fundamental question: can AI capability be purchased, or must it be grown?

The historical evidence is mixed. South Korea and Taiwan built world-class semiconductor industries through sustained state investment, technology transfer, and institutional development — but the process took decades. The Gulf states’ own experience with economic diversification has been uneven: the UAE’s success with Emirates airline and Dubai’s logistics hub demonstrates that strategic investment can build world-class capability in capital-intensive industries, but efforts to build knowledge economies have had more limited success.

AI may be different from previous technology waves in ways that favor the Gulf strategy. Foundational models can be licensed or partnered with rather than built from scratch. Cloud infrastructure can be purchased and deployed at scale. Inference (running AI models) is more capital-intensive and less talent-intensive than research (creating new models), and the Gulf states may find a viable niche as inference infrastructure providers even if they never produce frontier research.

Or the Gulf AI bet may follow the pattern of previous Gulf megaprojects: enormous initial investment, impressive physical infrastructure, and a persistent gap between hardware capability and the human expertise needed to use it. The data centers may be built. The GPUs may be installed. And the models running on them may be designed, trained, and maintained by American companies, with the Gulf states functioning as very expensive landlords rather than genuine AI powers.

INHUMAIN.AI will track the answer as it unfolds. The Gulf states’ AI experiment is too consequential to ignore and too large to fail quietly. Whatever happens, it will shape the global AI landscape for decades.

For broader geopolitical context, see: AI Geopolitics: Who Controls Inhuman Intelligence Controls the Century.

For tracking of Gulf-Western technology partnerships and their implications, see: HUMAIN Watch.