Home Innovate Playing The Long Game: Hassan Alnaqbi, CEO, Khazna Data Centers

Playing The Long Game: Hassan Alnaqbi, CEO, Khazna Data Centers

In a conversation with Inc. Arabia, Alnaqbi explains how the UAE-based hyperscale data center provider he leads is powering the region’s artificial intelligence ambitions.

Engy Ahmed
images header

The Middle East’s artificial intelligence (AI) ambitions are growing— and Khazna Data Centers is building the infrastructure to match.

In June, the UAE-based hyperscale data center provider announced a partnership with tech giant NVIDIA to build “AI factories” across the Middle East and Africa, positioning itself at the core of the region’s next technological leap. At the heart of the deal is the certification of Khazna’s next-generation data center design to support NVIDIA’s Blackwell architecture—the graphics processing unit (GPU)-heavy infrastructure purpose-built for the era of generative AI (gen AI), real-time digital twins, and autonomous systems.

But the move is about more than just tech specs. For Hassan Alnaqbi, CEO of Khazna, it’s a signal of his enterprise's broader ambitions to lead not only in infrastructure, but in shaping the region’s AI ecosystem. “This collaboration positions Khazna as a key regional and global player in the AI infrastructure landscape, demonstrating our ability to support high-density, high-performance compute environments at scale,” Alnaqbi tells Inc. Arabia. “It reflects the growing international confidence in our capabilities, and our strategic intent to help shape the future of AI-ready infrastructure not just in the UAE, but across emerging and established markets alike.”

Founded in 2012 by Abu Dhabi’s sovereign wealth fund Mubadala, Khazna operated as a national infrastructure player before being fully acquired by Abu Dhabi-based tech group G42 in 2020—a move that marked a turning point in its scale, ambition, and alignment with the UAE’s broader AI strategy. Since then, the company has rapidly evolved into one of the region’s most aggressive developers of hyperscale, AI-ready data centers, with AI clusters being designed for up to 250 megawatts (MW) per site, and future data halls reaching 50 MW each. As a part of G42, Khazna operates within what Alnaqbi calls “one of the world’s most advanced AI ecosystems.” Yet, its role is deliberately focused. “We’re not positioning ourselves as an accelerator or incubator,” Alnaqbi says. “We’re focused on building the foundations. The innovation that follows is something we’re proud to support—even if it’s not something we directly control.”

Playing The Long Game: Hassan Alnaqbi, CEO, Khazna Data Centers

Khazna's AUH6 Data Center.

NVIDIA’s certification of Khazna’s next-generation data facilities to support its Blackwell architecture, which is engineered for GPU-accelerated workloads, will fulfill a critical need as AI systems scale in both complexity and demand, explains Alnaqbi. “Blackwell-certified infrastructure is purpose-built for high-density, high-performance AI workloads, from training massive language models to running real-time inference for enterprise-grade applications,” Alnaqbi explains. “We see this partnership not as a finish line, but as a step forward, one that underscores the importance of collaboration across government, industry, and technology leaders to build the infrastructure that AI will rely on for decades to come.”

For Khazna, the partnership with NVIDIA—which came on the heels of US President Donald Trump’s visit to the UAE in May and the announcement of the five-gigawatt (GW) UAE-US AI campus—is more than a milestone. It is a step towards future-proofing AI infrastructure in a region eager to lead the next wave of digital transformation. “By investing in Blackwell-certified environments and working alongside world-leading technology partners like NVIDIA, we’re helping lay the physical and operational foundations the country needs to lead in the AI era,” Alnaqbi notes.

Another of Khazna’s partnerships took shape in July when it and Italy-based energy company Eni signed a head of terms to establish a joint venture to develop an AI data center campus with a total information technology (IT) capacity of 500 MW in Ferrera Erbognone, Lombardy. The project marks the first step in a broader program targeting a total IT capacity of up to 1 GW in Italy, as outlined in a letter of intent signed during the UAE President HH Sheikh Mohamed Bin Zayed Al Nahyan’s state visit to Italy in February.

The partnership also demonstrates the scale of Khazna’s ambitions, which has an expansion roadmap that stretches across France, Italy, Saudi Arabia, Türkiye, Egypt, and Kenya—part of a push to reach 1 GW in capacity. “Our expansion strategy is guided by a focused set of criteria: proximity to rising AI demand, alignment with national digital agendas, availability of critical infrastructure, and the presence of partners who understand the nuances of hyperscale delivery,” Alnaqbi explains, noting that the company prioritizes markets that are seeing rising AI adoption but limited local infrastructure to support it. “These are often regions where cloud penetration is rising, regulatory frameworks are maturing, and there’s a strategic interest in becoming regional, digital, or AI hubs,” he adds.

A scene from the announcement of the partnership between UAE-based Khazna and Italy-based EniA scene from the announcement of the partnership between UAE-based Khazna and Italy-based Eni.  

But Alnaqbi notes that scaling across diverse geographies comes with its own set of complexities. “Designing hyperscale AI infrastructure in new geographies inevitably involves complexity, from power availability and grid reliability, to climate adaptation and workforce readiness,” he shares. “However, we don’t see these as hurdles so much as design parameters we plan for from day one. We address them by adapting our modular, scalable build approach to local realities, by integrating sustainable technologies early, and by leveraging operational best practices we’ve refined at home in the UAE. Our teams engage early with regulators and utilities to de-risk timelines and ensure alignment with long-term national priorities.”

And as demand for AI compute takes shape across industries, Alnaqbi points to several use cases that could utilize Khazna’s facilities. “In the near term, we expect to see clients using our facilities for use cases such as autonomous systems, large-scale generative AI, AI-enabled cybersecurity, and real-time digital twins,” he says. “These are things that demand not only GPU muscle, but also consistent power delivery, thermal management, and high-speed interconnects.” Notably, Alnaqbi is seeing AI uptake transforming industries such as finance, energy, and manufacturing. “Financial services are among the earliest adopters, driven by demand for real-time analytics, fraud detection, and AI-powered customer engagement,” he says. “Energy and utilities are starting to use AI to optimize grids, reduce downtime, and accelerate sustainability efforts. And manufacturing and logistics can leverage AI for automation, predictive maintenance, and smarter supply chains. especially as the region builds new industrial capacity.”

While Khazna’s core focus remains serving large-scale enterprises, governments, too, are leaning in, embedding AI into smart city frameworks and citizen-facing digital services. As for startups, while Khazna’s infrastructure has been primarily tuned to hyperscale needs, Alnaqbi shares that the company’s capabilities are becoming relevant to AI startups scaling beyond the experimentation phase. “Fast-scaling AI companies leveraging large language models (LLMs) or building inference-heavy applications can look to us as a strategic partner when their compute demands outgrow public cloud economics or when data residency, latency, or performance become differentiators,” Alnaqbi says. “We’re not actively targeting early-stage incubation. But for startups ready to scale real AI products, especially across the MEA, we can offer foundational infrastructure that matches ambition with capability.”

And as the pace of change accelerates and demand shifts, Khazna is ensuring that its AI-ready infrastructure continues to be modular, agile, and constantly evolving. “We’re moving from static, one-size-fits-all facilities to agile, high-density environments built for the scale and complexity of modern AI workloads,” Alnaqbi says. “These future environments will need to deliver far more compute power, with greater flexibility and lower latency. They’ll need to support constant evolution—not just in terms of hardware, but in how space, power, and cooling are configured and optimized. And they’ll need to do all of this sustainably.” Alnaqbi also highlights that Khazna is designing for a range of AI workloads, which can go anywhere from LLM training and inference to generative AI pipelines and real-time analytics—thereby effectively building future-proofed infrastructure. “We don’t need to know the model being trained,” Alnaqbi adds. “We just need to ensure that whatever’s coming next— whether it’s Blackwell-scale LLMs or multi-tenant inference clusters—can run securely, efficiently, and at scale in a Khazna facility.”

KhaznaKhazna's Dubai headquarters.

And one of the keys to powering Khazna’s future-proofed infrastructure, Alnaqbi tells us, is developing an infrastructure that’s not only scalable but also has sustainability built into it. “Sustainability is not an optional extra for us,” he points out. “It’s a core design principle embedded into every stage of our roadmap.” As an example, Alnaqbi points out that wherever possible, clean energy is being integrated into Khazna’s developments. “The other side of the coin is efficiency with the power you are using,” he adds. “AI workloads generate significantly more heat than traditional enterprise compute, so we design with the most advanced cooling technologies that we can find. We have a whole innovation team dedicated to this, and their job is to balance performance with efficiency. Our average power usage effectiveness (PUE) of 1.4, better than the global industry average, shows how successful they’ve been. We’re particularly proud of this PUE figure—despite the fact that we’re operating in the harsh climate of the Gulf, we’re able to match or beat efficiency numbers seen in cooler regions.”

Beyond building infrastructure, Alnaqbi shares that Khazna is also helping nurture the broader AI landscape by investing in talent, operational expertise, and institutional partnerships to support the ecosystem that will power and maintain these facilities. “We’re already investing in talent development through dedicated programs focused on operational excellence, AI capability building, and partnerships with institutions like the Uptime Institute [a global digital infrastructure authority based in New York],” he says. “This ensures that, as AI infrastructure scales, so too does the local expertise required to run and evolve it.” And while Khazna may be a forerunner in the industry, Alnaqbi tells us that it hardly wants to be a lone provider. In fact, he points out that the vast opportunity around AI infrastructure in the region is simply too large for any single provider to tackle alone. As such, he stresses the importance of strategic partnerships, shared innovation, and collective efforts to push sustainability and performance standards forward. “These are not speculative bets,” Alnaqbi says. “Done right, AI-ready data centers are cash-generative infrastructure assets. They require upfront capital, yes, but they’re also financeable, and increasingly critical to national digital agendas.

However, this opportunity, Alnaqbi stresses, comes with a responsibility—as well as a commitment to long-term results. “If you’re building AI, you’re building for the core of the future economy,” he points out. “That requires not just infrastructure, but commitment, credibility, and the willingness to play a long game.”

Pictured in the lead image is Hassan Alnaqbi, CEO of Khazna Data Centers. All images courtesy Khazna Data Centers.

Reading time: 11 min read
Last update:
Publish date: