A UAE-based sovereign AI infrastructure and platform company is building one of the world’s most powerful sovereign artificial intelligence systems after securing a major deal with Nvidia for thousands of next-generation chips.
The move could strengthen the UAE’s control over its most sensitive data and AI systems, reducing reliance on global technology platforms.
Aleria announced last week that it will deploy 8,640 Nvidia Blackwell Ultra Graphics Processing Units in the US, with plans to expand to 16,000, while the UAE will receive early delivery of Nvidia’s DGX Vera Rubin systems, among the most advanced supercomputing platforms used to train and run AI models.
Why is the most advanced chip needed? Eric Leandri, chief executive of Aleria, believes it is necessary for the UAE to be truly sovereign – independently capable of developing, installing and governing its own AI.
“You need to have the best GPU in the world … if you want to run the best model,” Mr Leandri told The National in an interview. And Nvidia - with whom the firm has been working with for about five months since Aleria's inception in January 2025 - is far ahead of the rest, he added.
“When you want to do a sovereign AI you need to have full control, and a platform that can compete with the top platforms of the world,” he said. That you do not have to rely on other platforms.
He said if UAE SAI systems are competitive with public cloud providers such as OpenAI, Microsoft Azure and Google Cloud, they will not have to hand over their valuable information or intellectual property.
There’s a simple rule, he said. “Intelligence in the cloud … is the intelligence of somebody else.”
Mr Leandri explained that the UAE can achieve that level of competitiveness this year with the right investments.
“Instead of waiting for others to come and explain to us how to do it, we can do it here. Everything is made in the UAE.”
The 30% that matters
Despite the scale of investment, Mr Leandri says that sovereign AI does not need to replace global systems entirely.
“Seventy per cent of what you're doing does not need sovereignty,” he said, giving examples such as buying coffee, online shopping, or food delivery.
Instead, the focus is on the most sensitive 30 per cent of digital interactions with sensitive data including health care, financial systems, government operations and personal data, where control is critical.
“It's very important … for the countries, for our lives,” he said, especially during times where security is needed, such as the Covid pandemic, or currently amid the US-Israel war with Iran.
The effect of SAI is already being felt across industries. One example given by Mr Leandri is Abu Dhabi developer Aldar, where AI has significantly reduced the time required to design new residential projects – resulting in immediate cash savings.
“Months and months of work – now it's two days or one week,” he said, with such gains translated directly into cost savings and faster project delivery, particularly in sectors where planning and design are resource-intensive.
Cost of falling behind
Mr Leandri warned that failing to build sovereign AI capabilities could leave countries and companies exposed, not just technologically, but economically and strategically.
He said real-time AI systems are already transforming how organisations operate, enabling faster decision-making and deeper insights. But control over data alone is not enough.
“Data is a small part of the story,” he explained, and so is the physical hardware such as the Nvidia chips when discussing true sovereignty.
“You need to have the software control," Mr Leandri said, to dictate how all the data and technology will behave, securely within a closed system.
That philosophy underpins Aleria’s role within Abu Dhabi’s wider AI strategy. The company was developed inside International Holding Company, one of the UAE’s largest conglomerates chaired by Sheikh Tahnoon bin Zayed, also Deputy Ruler of Abu Dhabi, as part of a broader push into sovereign AI and strategic technologies, while also expanding in the US through large-scale GPU deployments and data centres.
Mr Leandri has spent decades advocating for greater control over digital infrastructure. As former head of the French search engine Qwant, he was involved in regulatory battles with Big Tech that led to multibillion-dollar antitrust penalties against Google.
At the core of his argument is the risk of handing over sensitive information, from intellectual property to investment strategies, to global technology companies.
“You have to keep it with you … their IP, their data, their investment, where they want to invest, what they want to do,” he said. “Because if everybody knows, then you will be in competition with these big groups."
Shifting global balance
At a broader level, Mr Leandri believes AI will reshape global power dynamics, with countries competing not just on technology, but on control of infrastructure and data.
The global race for sovereign AI is increasingly defined by three poles, he said. The US leads in sheer computing power and infrastructure, driven by companies such as Nvidia and hyperscale cloud providers, while China is building a self-sufficient ecosystem supported by a vast domestic market.
The UAE, by contrast, is emerging as a leader in sovereign AI deployment, focusing on controlling critical data and systems rather than competing on scale. As Mr Leandri put it, the US will remain “on top of the wave” due to its technological dominance, while other countries must build their own capabilities to retain control over their most sensitive digital assets.
A group of emerging middle powers, including Saudi Arabia in the Gulf, the UK and France in Europe, and India and South Korea in Asia, are also accelerating investments in sovereign AI to avoid dependence on global technology platforms.
As the technology evolves, the balance of power may shift towards countries that can build and control their own AI infrastructure, said Mr Leandri. The Nvidia deal represents a step towards that goal, providing the foundation for scaling sovereign AI from early deployments to full national capability.
Aleria is also preparing to expand its offerings further, including new AI applications designed to interact directly with users.
Mr Leandri said the next phase of AI will shift from tools to autonomous agents that act on behalf of users, carrying out tasks such as managing finances, analysing spending, and advising on major decisions like purchases or loans. Rather than simply providing information, these systems will connect to personal data from bank accounts to daily activity and execute customized actions in real time through a user’s device.
“We will release something like this in the next couple of weeks or months,” Mr Leandri said.
For governments and organisations navigating an increasingly uncertain global landscape, the message is clear: control over AI is no longer optional.
“Sovereignty first is not nice to have. It's a must have,” he said.
