The case for Gulf ‘free data zones’

How can corporations compete globally if American AI is more expensive than China?

February 20, 2025, 3:00 PM

By offering data sovereignty, the Gulf can make its data centres appealing to the global market

“Drill, baby, drill”, President Donald Trump bellowed during his inaugural address last month. But given the lacklustre oil price and the booming demand for data centres, perhaps he’d be better placed revising his catchphrase to “Build, baby, build”.

At present, there are 10,000 data centres worldwide and more than half of those are in the US, according to the Visual Capitalist website. Management consultant McKinsey says the data centre market will be worth $587 billion by 2026 – and global demand for the facilities could triple by 2030.

However, this global demand is not yet fungible or interconnected, because of strict data sovereignty laws in most countries.

Data centres are increasingly supporting artificial intelligence applications. AI data flows are something like the tech equivalent of Super 98, a high-performance but highly demanding fuel. As a result, these centres require significantly more energy and specialised cooling technology to prevent overheating.

In the past 24 months, typical data centre power consumption has almost doubled from 17 to 30 kilowatts per rack, McKinsey says. Such colossal usage means energy cost is a major factor in selecting centre locations.

The issue of data sovereignty raises challenges for companies and countries that want to squeeze down their AI costs

Microsoft, for example, is so focused on meeting its US data centre energy demands that it is working to restart the nuclear reactor at Three Mile Island in Pennsylvania.

Another crucial consideration when selecting data centres is the nature of the data being processed. Privacy, security and sovereignty laws dictate that personally identifiable information related to citizens cannot be transferred offshore in most jurisdictions.

While these protections benefit citizens, they also create challenges.

Demand for data centres in one country cannot simply be offset by surplus capacity in another. Currently, it’s not possible to partition sensitive personally identifiable information data from general data, so most national data is trapped in its sovereign land.

This issue of data sovereignty raises challenges for companies and countries that want to squeeze down their AI usage costs and, therefore, boost global competitiveness.

There’s clearly a gap in the market for an innovative location that can offer low-cost power, abundant land and a free trade zone operating environment for data centres. Could that be the winning formula?

This is where I see that the Gulf, blessed with cheap energy, could become the data centre capital of the world.

The UAE, for example, has extensive experience operating free trade zones across its seven emirates. These zones are set up to attract investment and offer benefits such as the possibility of 100 percent foreign ownership, tax exemptions and dedicated infrastructure.

But a further step is required. Beyond allowing full foreign ownership, data centre sovereignty can be granted to other nations – just like embassies. Such an offering would entice hyperscalers – such as Google, Amazon Web Services, Azure and Baidu – to establish data centres in the Gulf.

In 2021, the UAE became the first country in the region to implement mandatory AI regulation in one of its free zones, the Dubai International Financial Centre. There is potential to leverage competitive advantage by further advancing regulations and establishing a dedicated data centre free trade zone – or a “free data zone”– that could unlock significant value.

The GCC data centre market is forecast to rise to $17.1 billion by 2033, with the UAE and Saudi Arabia leading the race. If the region were to offer free data zones, this sum could be greatly amplified.

To future-proof the market, the centres must be designed to process any data, regardless of ownership or control. This shift would change the landscape, as nations would have the option to either build data centres domestically – likely at a higher energy cost – or construct them in a free trade zone to reduce operating expenses.

The AI boom is adding fuel to this fire. The geopolitical race for leadership requires investment in infrastructure such as data centres, power grids and water, not just Nvidia chips.

Consider President Trump launching the $500 billion Stargate Project to develop “physical and virtual infrastructure to power the next generation of advancements in AI”. This colossal venture, backed in part by Gulf investment, includes the construction of huge data centres.

Oracle, a founding Stargate member, plans to build 10 data centres at 500,000 square feet apiece, potentially expanding the global data centre footprint by 20 percent. But at what cost?

Expensive infrastructure leads to costly computing power, which inevitably gets passed on to the end user.

How can corporations compete globally if American AI is more expensive than its counterparts in countries such as China? AI supremacy isn’t just about being first or more accurate – it’s also about achieving those milestones cost effectively.

As the mathematician Clive Humby famously said in 2006, “Data is the new oil,” highlighting its potential to drive value through the insights it can provide when adequately refined and utilised.

With the rapid growth of AI, data has become more valuable than ever. However, its true worth will ultimately be determined by how efficiently it is stored and how freely it can move across silicon, wires and borders.

Nish Kotecha is a serial tech entrepreneur and former investment banker. He is the chair and co-founder of Finboot, chair of Agam.ai and advisory board of Innoviti