Gulf States Chase ‘Sovereign AI’ While Still Plugged Into American Hardware The pool of national engineers capable of building and sustaining sovereign AI is growing, yet remains limited By Jacob Wirtschafter/The Media Line [Istanbul] Across the Middle East, governments are racing to build artificial intelligence systems they can control, even as their visions of sovereignty […]
Science
The Media Line: Gulf States Chase ‘Sovereign AI’ While Still Plugged Into American Hardware
Audio By Carbonatix
Gulf States Chase ‘Sovereign AI’ While Still Plugged Into American Hardware
The pool of national engineers capable of building and sustaining sovereign AI is growing, yet remains limited
By Jacob Wirtschafter/The Media Line
[Istanbul] Across the Middle East, governments are racing to build artificial intelligence systems they can control, even as their visions of sovereignty still depend on American-made silicon.
At a technology fair in Dubai last month, a developer typed two Arabic words into a computer: “National Day celebration.” Within seconds, the screen filled with images from the Emirates’ national story — pearl divers rising through cyan waters, skyscrapers erupting from the Gulf’s shoreline, and women in traditional dress waving UAE flags, a sharp break from the sand dunes and falcons that Western image generators usually produce.
The software was Imagine&, the UAE’s answer to image-generating AI. Built by e&, the state-run telecom conglomerate turned tech giant, the tool is trained not on the diffuse swirl of global internet culture but on carefully selected Emirati imagery and national iconography.
Imagine& sits inside a broader rush toward digital sovereignty, the idea that a nation’s data, cloud architecture, AI models and computing infrastructure should remain under domestic control, guarded as tightly as borders or oil fields.
“AI is a state of sovereignty,” UAE Economy Minister Abdulla bin Touq Al Marri said. “Like you spend budgets on defense, on cyber, you have to spend on AI.”
The UAE has moved aggressively to position itself as a global AI hub. Abu Dhabi rolled out an Arabic-language model earlier this year designed to capture Gulf dialects. E& has deployed more than 160 machine-learning systems internally and partnered with Oracle to install NVIDIA H100 clusters in its data centers. According to Microsoft’s AI Diffusion Index, the UAE now leads the world in per capita generative AI use, with nearly 60% of the population using such tools.
Behind the branding lies a core strategic goal: keeping Emirati data inside Emirati borders. This year, e& and Amazon Web Services launched the UAE Sovereign Launchpad, a domestically controlled cloud intended to prevent foreign jurisdictions, including allies, from accessing sensitive information.
Washington has taken notice. In November, Microsoft became the first company under the administration of President Donald Trump to receive export licenses for advanced NVIDIA processors to the UAE, as many as 500,000 chips annually through 2027. At the same time, Microsoft pledged more than $15 billion in new investments before the decade’s end.
“AI is not just about code,” UAE Energy Minister Sultan Al Jaber said. “It’s about gigawatts.”
Every advanced chip draws as much electricity as a suburban home, and the Emirates, flush with low-cost power, see a natural advantage.
Yet as the infrastructure expands, a different issue has grown more sensitive: who is allowed to speak publicly about the AI powering the Emirati vision and who is not.
Despite marketing Imagine& as “Emirati-led,” e& declined multiple requests to provide a single Emirati engineer or spokesperson for this article.
For Marc Owen Jones, associate professor at Northwestern University in Qatar, the silence says a lot.
“Many Emirati professionals avoid visibility because unscripted comments can be interpreted as a political signal,” he told The Media Line. “The state promotes an image of Emirati-led innovation, yet the people who could give that image credibility often stay silent out of self-protection.”
Jones said Gulf governments’ talk of sovereign AI often masks locally branded interfaces built atop Western models that still rely on US chips and infrastructure. Any claim to independence, he argued, remains fragile so long as national strategies depend on uninterrupted American export approvals.
The UAE’s labor force stands at roughly 9.4 million people, but nearly 90% are expatriates. The pool of national engineers capable of building and sustaining sovereign AI is growing, yet remains limited. Saudi Arabia faces the same problem: Despite large-scale investments in training programs, the kingdom still reports a 50% hiring gap in advanced AI roles.
Turkey and Israel, by contrast, enter this competition with deep homegrown talent.
Turkey’s information and communications technology sector employs more than 213,000 workers, supported by nearly 100,000 science, technology, engineering, and math graduates each year and more than 1,200 research and development centers. Israel’s tech workforce exceeds 360,000 people, more than 10% of its labor market, the highest concentration of tech workers per capita in the world.
Both countries, though, face internal pressures that could weaken that edge.
Turkey’s rapid expansion of İmam Hatip religious schools, with enrollment surging from 65,000 students in 2002 to 1.3 million today, has raised concerns that math, science, and engineering are losing ground. In Israel, most ultra-Orthodox schools still avoid core subjects, and only about 20% of Haredi men work in STEM-related fields.
Then came Oct. 7.
Following the attacks and the Gaza war, surveys by the Israel Innovation Authority found that 15% to 20% of Israeli tech workers were considering relocation during 2024–25. More than 100,000 Israelis left the country temporarily or permanently during the first year of the conflict, many from the high-tech sector. Some firms are already shifting talent to the Gulf: The Israeli outsourcing company AllSTARSIT plans to relocate 250 developers to Dubai in one year and 750 within two.
Turkey faces its own exodus. More than 150,000 skilled workers emigrated during 2023–24 amid inflation, currency volatility, and shrinking academic and press freedoms.
If the Gulf casts AI as cultural stewardship, Turkey casts it as a matter of defense.
At a technology forum in Istanbul, Erkam Tüzgen, head of Turkey’s Bilişim Vadisi technology park, called his institution “the command base of Turkey’s technology and innovation ecosystem.” The park hosts more than 700 firms, many working on autonomous systems, encryption, and military AI.
Baykar, Turkey’s flagship drone manufacturer, now claims 65% of the global unmanned aerial vehicle market. This fall, the Bayraktar TB3 naval drone completed more than 100 fully autonomous sorties from the TCG Anadolu using domestically developed AI — a world first for carrier-based drone systems. Serial production has begun for the Kızılelma, Turkey’s first jet-powered unmanned combat aerial vehicle, expected to enter service in 2026.
In November, Baykar integrated the TOYGUN low-observable targeting system into the Kızılelma — technology previously found only on fifth-generation fighters. A month earlier, Baykar’s AI-guided KEMANKEŞ-1 mini cruise missile struck airborne targets in live-fire tests from an Akıncı drone, marking the first recorded use of such a system.
Ambitions extend well beyond Turkey’s borders. In September, Italian defense giant Leonardo signed a joint venture with Baykar to co-develop drones, with the Kızılelma under consideration as a “loyal wingman” for the sixth-generation fighter being developed by the UK, Japan, and Italy.
“Digital power becomes both shield and signal,” said Erman Akıllı, a researcher at Hacı Bayram Veli University. “It fortifies national resilience against asymmetric threats and signals Turkey’s readiness to contribute to a more stable and equitable world order.”
Yet the sovereignty project has limits. The Kızılelma runs on a modified Ukrainian-designed AI-322F turbojet, revealing how even Turkey’s most advanced systems still depend on foreign components. And despite rapid adoption among individuals — 19.2% of Turks now use generative AI — only 7.5% of Turkish businesses report AI integration. A draft AI regulation bill introduced in 2024 remains stalled in parliament.
Saudi Arabia is pursuing the most ambitious and expensive sovereignty strategy. During Crown Prince Mohammed bin Salman’s recent visit to Washington, the kingdom announced $1 trillion in planned U.S. investments. At the center sits HUMAIN, a new AI company backed by the Public Investment Fund.
HUMAIN plans to build up to 6 gigawatts of data center capacity by 2034 and is partnering with NVIDIA, AMD, Cisco, Amazon Web Services and Elon Musk’s xAI. A single 500-megawatt HUMAIN–xAI data center could house up to 600,000 NVIDIA GB300 graphics processing units — one of the world’s largest planned clusters.
Michael Ratney, former US ambassador to Saudi Arabia, said during a Nov. 17 AJC webinar that the kingdom’s strategy remains rooted in Washington. “Saudi Arabia wants a relationship with the United States that is fundamentally anchored and that lasts for decades,” he said.
Yet more computing power also means more potential for control: “With AI, governments can tailor culture itself to what they believe is most efficient for them,” Henna Anuksha Varra, digital rights advocacy fellow at Americans for Democracy & Human Rights in Bahrain, told The Media Line. “They can get that message out everywhere.”
Her colleague James Suzano, Americans for Democracy & Human Rights in Bahrain’s legal director, said the UAE’s surveillance systems are already operating “at scale,” with smart-toll grids and license-plate readers feeding data into citywide monitoring networks. The use of spyware against dissidents “isn’t theoretical,” Suzano said, citing the case of imprisoned Emirati activist Ahmed Mansoor.
“AI doesn’t create surveillance,” Varra added. “It accelerates it.”
“For Gulf states, tech independence still depends on the United States,” Elizabeth Dent, a fellow at the Washington Institute for Near East Policy, told The Media Line. “Even if they become more independent, it’s built on US infrastructure.”
China offers another path, through Huawei, Alibaba Cloud, and Dahua, with no political conditions attached.
A US official put the dilemma bluntly: “If we reject them, we drive them into China’s arms.”
For all the investment and political theater around digital sovereignty, the region faces a basic constraint: None of its powers produce advanced chips, and no regional AI model rivals GPT-4 in scale.
That leaves a quiet unease at the heart of the sovereignty push, the fear that it can be performed or branded, but not fully achieved.
As Varra put it: “The very loud effort to build a national AI brain suggests anxiety that AI is a threat to their national brain.”