The AI giant is reportedly backing a 5-gigawatt data centre in Abu Dhabi, a facility so vast, it would dwarf Monaco and could reshape global AI infrastructure.
A Bold Expansion Into the Gulf
OpenAI, the company behind ChatGPT, is reportedly helping to develop one of the largest AI data centre campuses on the planet, a 10-square-mile megaproject in Abu Dhabi that could eventually outsize the entire nation of Monaco.
The plans (first reported by Bloomberg) describe a 5-gigawatt facility that would draw as much power as five nuclear reactors and become a key node in OpenAI’s growing Stargate infrastructure programme. The UAE site is being built in partnership with G42, a powerful Abu Dhabi-based tech group chaired by the country’s national security advisor, Sheikh Tahnoon bin Zayed Al Nahyan.
Sources familiar with the project say OpenAI would act as one of the primary anchor tenants, though final details are still being worked out. If confirmed, it would mark OpenAI’s largest international infrastructure commitment to date, and one that dramatically raises the stakes in the global race for AI compute power.
Why Abu Dhabi?
For OpenAI, the UAE offers more than just open desert and deep pockets. The region has become a key strategic ally in the US-led effort to expand AI capacity without ceding ground to China.
Abu Dhabi, in particular, has poured billions into tech development through sovereign funds and national champions like G42 and MGX. In return, it has sought deeper technological integration with major US firms, a goal reinforced during President Trump’s recent Middle East visit, where AI cooperation featured prominently on the agenda.
In fact, this new data centre forms part of a wider bilateral framework signed in May between the US and UAE to accelerate AI deployment across the Gulf. The accord could include the sale of over a million high-end Nvidia chips to the UAE, a move intended to counter Beijing’s influence in the region.
OpenAI CEO Sam Altman has long praised the UAE’s ambition. For example, speaking in Abu Dhabi last year, he said the country “has been talking about AI since before it was cool.” His company’s relationship with G42 dates back to 2023, when the two announced a regional partnership to promote AI adoption across the Middle East.
The Global AI Infrastructure Play
The Abu Dhabi build looks like being the most ambitious element yet in OpenAI’s “Stargate” initiative, i.e. a joint venture with SoftBank and Oracle aimed at rolling out hyperscale AI infrastructure around the globe. The programme, unveiled in January alongside former President Trump, is expected to channel up to $500 billion into global projects over the next four years.
OpenAI’s first Stargate campus is already under construction in Abilene, Texas. That site is expected to reach 1.2 gigawatts, making the planned Abu Dhabi facility more than four times larger! For context, even Microsoft’s largest data centres currently consume less than 1 gigawatt.
It’s been reported that each site will serve as a supercharged AI hub, equipped with high-density computing clusters and the latest generation of Nvidia and AMD chips. These are designed to handle the massive workloads required for training and deploying next-gen language models, vision systems, and multimodal tools.
According to OpenAI, demand for compute is growing faster than chipmakers can supply. For example, speaking earlier this year, Altman warned that “AI progress will soon be bottlenecked not by algorithmic innovation, but by energy and silicon.” It seems, therefore, that Stargate is OpenAI’s attempt to remove that bottleneck.
Power, Partnerships, and Politics
The scale of the Abu Dhabi site is breathtaking. For example, at full capacity, it would:
– Span over 10 square miles, which is larger than many cities and territories, including Monaco (0.78 square miles).
– Consume 5 gigawatts of power, which is equal to around 5 large nuclear plants or roughly the output of the UK’s Hinkley Point C when complete.
– House an ecosystem of tenants such as OpenAI, Oracle, and potentially others, although exact names have not yet been disclosed.
Oracle, which has been expanding its cloud footprint aggressively to support AI workloads, is reportedly involved in developing the first phase of the campus. Microsoft, OpenAI’s biggest backer, is also entangled via its recent $1.5 billion investment in G42. Microsoft President Brad Smith has since joined the G42 board.
Complex Politics
However, the politics behind these partnerships are complex. For example, G42’s historical ties to Chinese firms, including Huawei and the Beijing Genomics Institute, have caused alarm in Washington. In 2023, US lawmakers warned that such links could enable Chinese access to sensitive American tech via the back door.
Under pressure, G42 has since claimed to have exited all China-related investments and operations. For example, as CEO Peng Xiao said (to Bloomberg) in January: “All of our China investments that were previously made are already divested. Because of that, we have no need anymore for any physical China presence.”
That change appears to have helped pave the way for closer US-UAE collaboration. That said, it seems that not everyone in Washington is convinced the threat has vanished. For example, some Trump administration officials are said to be uneasy about the symbolic and strategic risks of moving advanced AI infrastructure offshore, particularly to a region still watched closely by Chinese and Russian intelligence services.
Changing The AI Landscape?
If built as planned, the Abu Dhabi super-campus would not only change the geography of AI infrastructure, but could reshape how companies access and deploy AI globally.
For OpenAI’s users, especially enterprise clients, the move signals two big trends:
1. More global redundancy and resilience. Spreading infrastructure across continents could make OpenAI services more stable, scalable, and compliant with local regulations, including data residency requirements.
2. Faster innovation cycles. With massive compute on tap, OpenAI can continue training ever-larger models, rolling out more powerful tools for business, research, and defence applications.
At the same time, the project highlights the deepening alliances between Big Tech and geopolitical power blocs. As AI becomes a strategic asset akin to oil or semiconductors, access to compute may increasingly shape who leads (and who lags) in the next digital revolution.
A New Benchmark
For OpenAI’s rivals like Google DeepMind, Anthropic, and Meta, the Abu Dhabi development looks like setting a new benchmark. Few others have announced anything on this scale. It also raises questions about whether public cloud infrastructure alone will be sufficient to keep pace, or if more firms will start investing in their own AI megastructures.
Environmental and Other Concerns
Despite the hype around the scale and the promise of the project, it also raises tough questions about energy use, environmental impact, export controls, and the ethics of building frontier AI infrastructure in politically sensitive regions. These tensions are unlikely to go away anytime soon.
What Does This Mean For Your Business?
OpenAI’s reported potential leap into Abu Dhabi is more than just a headline-grabbing land grab or another tech tie-up. In fact, it appears to mark a pretty decisive moment in the evolution of AI infrastructure, i.e. where geography, politics, and compute capacity intersect on a global stage. If the deal goes ahead as anticipated, the Middle East may soon host the world’s largest AI campus, symbolising not just technical ambition but also a fundamental shift in where the power behind AI is physically rooted.
For OpenAI, the project signals a pragmatic move to secure the raw energy and silicon resources needed to train future generations of frontier models. The scale of the proposed facility underlines just how central compute has become to AI’s progress, and just how willing firms are to chase that capacity beyond their traditional home markets. Yet that ambition comes at a cost. The growing proximity between powerful AI companies and state-linked partners in geopolitically complex regions raises real questions about governance, transparency, and the safe stewardship of cutting-edge technologies.
UK businesses watching this from afar may be in two minds about its benefits. For example, on the one hand, expanded infrastructure could eventually translate into more stable, powerful AI services (and potentially lower costs) as scale economies kick in. For sectors already adopting generative AI at pace, such as finance, marketing, logistics, and law, that’s a tantalising prospect. However, it also reinforces the extent to which global supply chains, regulatory alignments, and geopolitical tensions now play an invisible but important role in the tools UK firms depend on.
Likewise, for policymakers, the Abu Dhabi campus could serve as a reminder of how fast the AI ecosystem is becoming a transnational infrastructure issue, one that affects everything from energy use and international trade to national security and data protection. The decisions being made in Abu Dhabi, San Francisco, or Washington are no longer abstract for UK stakeholders, but they shape the very foundations of how AI is accessed and governed globally.
In the end, the success or failure of this move may depend as much on diplomacy and ethics as on architecture and engineering. Although the vision is immense, so too are the risks. Whether this data centre becomes a beacon of innovation or a lightning rod for criticism will depend on how carefully all those involved, i.e. OpenAI, G42, Oracle, and the governments behind them, navigate the road ahead.
