Building AI's Backbone in the Middle East

Newsweek

a day ago

Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content.
When U.S. President Donald Trump announced the giant Stargate artificial intelligence venture was coming to the United Arab Emirates, data center company Khazna was already gearing up to become its backbone.
The shift to supporting AI, and the extra speed of implementation required, marked a massive scaling up from Khazna's cloud computing origins and would be a step towards delivering AI as a commodity in the UAE and beyond, Chief Executive Officer Hassan Alnaqbi told Newsweek in an interview in UAE.
"It's completely changed, completely changed. They need it now, maximum six months. And this is why I think my biggest challenge is developing Stargate in a record time," Alnaqbi said.
Majority-owned by the Abu Dhabi-based G42 holding company, Khazna has grown from a 28 megawatt company in 2021 to a half-gigawatt (500 megawatt) company now and is on course to meet the Stargate demands to provide five gigawatts (5,000 megawatts), with data centers being measured by their power consumption.
"I think the data center became sexy only after AI became the norm," he said. "It's part of the new intelligence grid, because you need energy, and of course AI, infrastructure, GPU, etc, but without the backbone, without the foundation layer, nothing can happen."
Stargate UAE will be operated by OpenAI and Oracle in a collaboration that also includes Cisco, Japan's SoftBank Group and NVIDIA, will supply its Grace Blackwell GB300 systems, after Trump lifted restrictions on advanced chip exports to UAE.
It is already one of the most connected societies in the world and prides itself on smooth delivery of digital services. Not only does it have a minister specifically for AI, but its National Artificial Intelligence System will formally become an advisory member of the cabinet as of next year.
Khazna CEO Hassan Alnaqbi
Khazna CEO Hassan Alnaqbi
Khazna
Going Global
Building the data centers, Khazna aims to be the foundation for AI operating far beyond UAE itself, both as a hosting itself and working with other countries to help build the necessary infrastructure. It already has projects in France, Italy and Turkey.
"Let's face it, I think UAE doesn't need the five gigawatt worth of GPU intelligence. It probably needs maybe let's say half a gigawatt, maybe even less, but the remaining is actually a platform for all the collective ecosystem building here to actually export it," Alnaqbi said.
"Where they don't have access to energy, where they don't have access to GPU, it becomes a service and the long term vision we have is that AI becomes a utility, just like electricity."
A long-term target is industrializing countries of the Global South that lack the UAE's level of infrastructure. Kenya and Egypt are already among those seen as strong prospects.
But closer to home is bigger neighbor Saudi Arabia.
"Saudi for us, is a very strategic market. It's bigger than UAE from all aspects, consumer industries, but also it's an underserved market, and I think with the whole division digitalizing and also now aggressively working and taking a leading position with AI," he said.
"We see ourselves as complementary to the government there in terms of coming with a track record and helping build that infrastructure."
Given that cooling is one of the biggest challenges in running a data center, and even more so for AI, one of the world's hottest countries might not be seen as an ideal place to be building it. The temperature averages near 100 degrees in summer, even if it is much more comfortable in winter.
At the same time, UAE has a goal of getting to net zero carbon emissions by 2050 as nuclear and solar become more important sources of electricity that is currently generated mostly by natural gas.
Khazna was leading innovation to design efficient data centers and cooling systems, Alnaqbi said. Despite the challenging environmental conditions, Khazna's power usage effectiveness (PUE) was already on a par with data centers in Europe, Alnaqbi said.
"We are also moving away from the air-cooling solution to the water-based cooling solution," he said. "We can save up to 70% of energy by utilizing the liquid to liquid cooling and on top of that, the water we're using for the cooling is also treated sewage water instead of drinkable water."
The expanding AI capabilities would themselves feed back into improving power usage, Alnaqbi said.
"I'm 100% sure not every company, not every industry is utilizing the power to the full potential," he said. "I can help you optimize that. This is the foundation layer of the new digital economy."