Billions in investments in major AI labs: where computing ends and surveillance devices begin

When the major Western AI labs opened their investment books to sovereign wealth funds, the scene at first appeared to be a massive liquidity flow to cover rising computing costs. But in Abu Dhabi, behind these partnerships stands an economic empire estimated at about $1.5 trillion, led by UAE National Security Advisor Sheikh Tahnoun bin Zayed Al Nahyan, placing himself at the heart of the global AI infrastructure. This rapid rise brings to the fore critical questions about the convergence of sovereign security systems with future algorithms.

At the core of this expansion, G42 emerges, a cloud computing and AI giant in the UAE. According to an investigative report published by The Guardian and carried by Middle East Monitor, the company was built in the late 2010s on the remnants of DarkMatter, which ran the secret espionage program known as Project Raven. The program initially incorporated former U.S. intelligence personnel and later expanded to target dissidents, journalists and human-rights defenders. Reports indicate that senior staff moved to the new entity, with G42’s current chief executive, Bing Xiao, having led a major unit at DarkMatter, while former company head Karim Sabagh now chairs one of the group’s subsidiaries.

The shift from digital surveillance software to behavioural-prediction tools moves influence from merely tracking devices to managing the entire cloud infrastructure.The company now markets predictive tools to governments aimed at analysing behaviour, mapping relationships and classifying suspects. The investigation also recounts documented instances of espionage technologies being used against dissidents, including engineer Ahmed Mansour, who was convicted in 2017 after a secret trial and is serving a sentence in Al-Sadr prison about ten miles from G42’s super-computing complex, as well as the testimony of British researcher Matthew Hedges regarding the conditions of his detention and his attribution of responsibility to the security-chain leadership. In response, G42 has officially denied the accusations, stating that it was founded as an AI and cloud-computing specialist and rejects any link to digital espionage or unlawful surveillance.

Despite these sensitive backgrounds, the company has enjoyed broad Western investment and political openness. Microsoft injected a $1.5 billion investment into G42, and the group and its investment funds have become partners and investors in major global firms such as OpenAI, Anthropic and XAI. On the U.S. government side, the Donald Trump administration lifted previously imposed restrictions and allowed the export of hundreds of thousands of advanced AI chips to the UAE, directing a large portion to G42’s data centres. The deal came two weeks after a Tahnoun-linked investment vehicle purchased $2 billion in cryptocurrency issued by World Liberty, a company tied to the Trump family in which it holds a 49 percent stake, prompting sharp criticism in Congress from legislators such as Elizabeth Warren on grounds of a conflict of interest, while the White House and the company denied any link between the two transactions.

For policymakers and tech leaders in the Middle East, whether in Gulf sovereign wealth funds or tech firms in Egypt and the Levant, this landscape forces a recalibration of infrastructure choices. Relying on local cloud services and models hosted within regional data centres is no longer a purely technical decision tied to inference efficiency or direct cost reduction; it has become a strategic decision linked to data compliance and sovereign classifications. Any geopolitical tension or potential tightening of Western export-control and surveillance rules on advanced chips in the region could place companies and institutions that build their critical systems on these shared platforms under sudden audit requirements and governance risks that exceed ordinary commercial contracts.

The battle for AI dominance no longer revolves solely around innovative algorithms; it has become a contest over who controls the data centres, advanced-hardware licences and the political levers that safeguard their flow.While sovereign wealth funds strengthen their presence at the core of global technology, the real challenge for regional institutions remains balancing the use of this massive computing power with protecting the sovereignty of their operational data.