Saudi Arabia and the United Arab Emirates dominate regional data center capacity and have mobilized substantial investments for AI infrastructure by 2030. It is the Middle East’s most visible bet on the digital economy, and perhaps the most ambiguous one. Building the pipes of global artificial intelligence does not guarantee mastering its production, or capturing its value durably.
The essentials
- Saudi Arabia and the United Arab Emirates dominate regional data center capacity and are mobilizing substantial investments in energy, networks, and AI infrastructure by 2030.
- According to the SDAIA, Saudi Arabia ranks 14th globally and 1st in the Arab world in the Global AI Index.
- The strategy mobilizes significant capital for infrastructure without necessarily creating conditions for endogenous innovation.
- Economist Philippe Aghion emphasizes the central role of competition in growth based on innovation and creative destruction.
- The signal to watch by 2030 is less the size of data centers than the emergence of local startups capable of producing AI applications.
Significant investments: the use that determines the scope
The mechanics are simple to describe. Gulf states have abundant capital, cheap energy, available land, and political will to accelerate economic diversification. AI infrastructure checks all the boxes: massive investment, international announcement effect, displayed diversification of revenues. Gulf states have announced substantial investments in data centers, high-speed networks, and energy capacity dedicated to the digital sector.
Behind the figure, the reality is more precise. Foreign hyperscalers are developing their presence alongside locally supported operators and projects. Saudi Arabia is indicated as 14th globally in the Global AI Index cited by the SDAIA. This ranking measures deployment speed, not ecosystem depth. Building fast, at scale, with public capital: it is exactly what petroleum rents had taught how to do.
Hosting the infrastructure of global AI presents distinct challenges from creating conditions for developing homegrown AI.
Aghion’s thesis applied to the desert
Philippe Aghion, in The Power of Creative Destruction, develops a central thesis: sustainable growth comes from within. It emerges when entrepreneurs challenge existing players, when markets are open to competition, when the state creates favorable conditions without substituting its allocation for that of the market. Without these mechanisms, without what economists call endogenous innovation, centralized capital allocation risks maintaining forms of dependence on central decisions.
Applied to the Gulf, the thesis is precise in its predictions. Gulf states allocate capital by sovereign decision to sectors chosen by planners. Foreign actors arrive, build, operate. Revenues flow into public coffers or sovereign wealth funds. But the local startups that could have challenged existing operators, the engineers who could have invented their own language models, the researchers who could have produced breakthroughs published in Nature or NeurIPS: they remain rare, and their rarity is no accident.
The environment matters as much as capital. A data infrastructure requires, to be embedded in a productive ecosystem, competitive research universities, labor markets attractive to talent, and a culture of entrepreneurial risk. The Gulf has understood part of the problem: KAUST in Saudi Arabia recruits international researchers, and Mohamed Bin Zayed University of Artificial Intelligence in Abu Dhabi is the world’s first university entirely dedicated to AI. These are real signals. But they remain modest compared to the scale of infrastructure investments.
The competing reading: the state as agent of catch-up
A different reading exists, and it deserves to be taken seriously. Mariana Mazzucato, in her work on the entrepreneurial state, shows that major technological breakthroughs—internet, GPS, semiconductors—have often emerged from large, directed public decisions, not spontaneously organized markets. In this perspective, the Gulf’s centralized allocation toward AI infrastructure reproduces less the petroleum rent than it prepares a foundation on which a local industry could rise.
The argument has real force. Experiences in technological catch-up show initial phases of massive public investment in basic infrastructure: energy, networks, training. South Korea in the 1970s–1980s, Taiwan and its foundry champions, Singapore and its digital port all began by building before innovating.
The state can prime the investment, but priming must lead to an ecosystem capable of generating endogenous innovation and competitive dynamics.
That is precisely where the two readings diverge. Aghion recognizes the state’s role in the initial phase; he insists on the necessity of a transition to contestable and competitive markets. Asian trajectories combine initial intervention and progressive opening to competition. Gulf states attract foreign giants as operators, while local markets remain concentrated on a limited number of players. Saudi and Emirati entrepreneurial ecosystems are dynamic and supported, while still presenting limits, notably in sustaining new enterprises.
Public contracts go to large integrators.
Lessons from the history of energy rents
The comparison with fossil rents is not rhetoric: it is a useful analytical structure. For five decades, Gulf states exported oil, imported skilled workers, and repatriated revenues to sovereign wealth funds. Gulf economies remain strongly tied to hydrocarbons and the state, while pursuing diversification strategies and private sector development. Local enterprises operate mainly in protected sectors—construction, commerce, public services—where petroleum rents create captive demand.
The AI strategy presents structural characteristics recalling dynamics of centralized resource allocation. Foreign companies participate in certain projects, but Saudi and Emirati local entities are also building data centers; global digital giants are partners or clients of certain projects, while local operators such as HUMAIN and Khazna also operate infrastructure. Value captured locally includes maintenance employment, tax revenues, and knowledge transfer, but remains concentrated in certain segments. The Gulf is significantly strengthening its role as host and investor in global AI, while also developing certain local production capacities.
Owning global digital infrastructure represents a substantial geopolitical and economic position. Several comparable analyses illuminate this point, including the energy transition documented by Daniel Yergin, which shows that infrastructure transformations of this scale take decades and follow nonlinear paths. An infrastructural rent remains a rent nonetheless: it finances the present without necessarily preparing what comes after.
The bets that would make a difference
Three levers could transform infrastructure into an innovation ecosystem. Each is partially engaged; none is advanced enough to conclude.
The first is research. KAUST and MBZUAI publish in the best international journals. Saudi Arabia has launched the Roshn Group and several Smart City platforms, which generate real data and potentially novel applied research questions. The number of AI researchers trained locally remains limited in international comparison, and the capacity to retain talent after training is an unresolved problem.
The second is the startup ecosystem. The tech startup scene in the Middle East is progressing; companies like Careem (acquired by Uber) or Anghami (music streaming) have shown that regional exit strategies are possible. But valuations remain modest, venture capital markets shallow, and a growing number of regional AI startups combine integration of foreign solutions with progressive development of local capacities. For a parallel on the difficulty of creating formal employment at scale in developing economies with heavy investment, the Indian example is instructive.
The third is regulation. The Emirates have a federal decree-law on personal data protection and their AI charter makes privacy, data security, and legal compliance explicit priorities, while maintaining openness to experiments in hopes of attracting players. It is a coherent bet. Its risk is symmetrical: lax regulation attracts experimenters, not necessarily long-term builders who need a solid rule of law and predictable markets.
The 2030 indicators
Saudi plans are structured around 2030, while the Emirates’ national AI strategy is commonly formulated to 2031; this horizon is both near and far. Near because data centers build fast, three to five years from project to operation. Far because an innovation ecosystem is measured in generations of trained researchers and entrepreneurs, not installed megawatts.
Two trajectories are plausible at this horizon. In the first, the Gulf becomes a solid regional hosting hub, attracts applied research centers from multinationals, and develops a local integration and deployment industry. Innovation remains imported, but value captured locally increases gradually. It is the scenario of digital Singapore, an economy that plays a central role in flows without controlling their sources. It is a defensible outcome.
In the second trajectory, infrastructure investments create conditions for a more autonomous ecosystem. Researchers trained at MBZUAI found startups that train models on high-quality Arabic data—language, health, agriculture, regional logistics. Venture capital markets deepen. AI applications designed in the Gulf export to Africa, South Asia, and the Arabic-speaking world. This scenario exists in strategic plans.
It demands decades, not billions alone.
What will allow distinguishing the two trajectories before 2030 is not the number of operational data centers. The relevant signals lie elsewhere: the number of AI publications by locally trained researchers in peer-reviewed international journals; the volume of venture capital raised by startups founded in the region without sovereign capital; the share of data center revenues that remains in the domestic economy rather than repatriated by foreign operators. These metrics are today largely absent from official announcements.
The April 2026 IMF calls on GCC economies to strengthen their diversification agendas and resilience to shocks, highlighting persistent needs for private sector development. AI infrastructure fits into this broader dynamic of progressive diversification.
A model to watch, not yet to imitate
The Gulf builds fast and at scale, an asset it deploys fully in digital infrastructure. For the coming decade, the central question is whether this building speed leads to innovation capacity or consolidates a position as top-tier host without producing conditions for endogenous creation.
Aghion’s thesis establishes that contestable markets are necessary to innovation dynamics. Asian technological catch-up experiences show variable combinations of public intervention, private capacities, and competition. In Saudi and Emirati plans, the state remains the central actor, the principal allocator and order-giver, which makes this withdrawal uncertain. It is the variable to watch.
Trajectories transforming an initial rent into a productive ecosystem have combined talent formation and environments favorable to entrepreneurial risk and experimentation. Infrastructure investments alone are insufficient. Decisions related to talent and human resource formation will be determining.
Sources
- S&P Global Market Intelligence & Oxford Economics, Key Themes Shaping Middle East & North Africa 2026 (January 2026): https://www.spglobal.com/market-intelligence/en/news-insights/research/2026/01/key-themes-shaping-middle-east-north-africa-mena-2026
- Philippe Aghion, Céline Antonin, Simon Bunel, The Power of Creative Destruction, Odile Jacob: https://www.odilejacob.fr/catalogue/journees-internationales/21-avril-journee-mondiale-de-la-creativite-et-de-l-innovation/pouvoir-de-la-destruction-creatrice_9782738149466.php
- IMF, MENA Economic Update, April 2026 (international monetary fund, regional report)
- Oxford Economics, GCC Economic Outlook 2026 (quarterly report, no stable URL)
