G42 chief: UAE aims to shift from energy exporter to intelligence exporter

The latest:

The UAE should reposition itself as a global exporter of intelligence rather than energy, G42 group chief executive Peng Xiao told the Dubai AI Festival on Monday. He said the country’s reliable energy supply and its speed in building computing infrastructure let it meet international demand, and that American partners are already coming to the region for compute capacity because it can be delivered now.

Details:

• The pitch: Xiao framed the shift as converting energy resources into computing power capable of supplying AI services far beyond the UAE’s borders, positioning that infrastructure as the Emirates’ next strategic export after hydrocarbons. He tied the ambition directly to the speed at which data centre capacity is being brought online.
• The build: G42 is developing a five-gigawatt AI campus in Abu Dhabi, Xiao said. It includes the one-gigawatt Stargate UAE project, with partners named as OpenAI, Oracle, Nvidia, Cisco and SoftBank. He did not give a completion date for the wider campus.
• The reach: Xiao said the UAE’s connectivity could allow services from the planned infrastructure to reach a potential user base of about four billion people. The figure was presented as addressable reach rather than committed customers, and no timeline was attached to it.
• On human work: Xiao argued that people should fundamentally rethink what they value, including the ability to think and use tools, because AI can now do both better. He said AI can already out-think humans, and urged a shift in value toward human connection and creativity.
• The ministerial warning: Omar Al Olama, Minister of State for AI, Digital Economy and Remote Work Applications, told the same forum that disruption is on the horizon for the private sector and for governments. He said the state must ensure every worker sees AI as an enabler and force multiplier rather than a displacer.
• Policy tempo: Al Olama said UAE leadership now reviews its AI plans every six months, up from an annual cycle. He traced the shifting focus from computer vision and self-driving cars in 2017, to large language models in 2022, and then to agents able to carry out tasks on a user’s behalf.
• The output goal: Al Olama said the aim is for a nation of around 10 million people to have the impact of one of 50 million or 100 million, by multiplying the output of everyone who works, lives in or visits the UAE through the technology.
• Government mandate: In April, Sheikh Mohammed bin Rashid, Vice President and Ruler of Dubai, launched a plan to deploy agentic AI across 50 per cent of federal government sectors within two years, covering services, operations and staff training. Al Olama said the expectation was to re-engineer processes from scratch.
• The caveat: Al Olama cautioned that success depends on how effectively organisations use AI, not how much computing power they consume. Companies must identify advantages rivals cannot easily copy, he said, warning that those relying on public models and cloud services risk being displaced or replaced.

Between the lines:

The two speeches point in different directions on the same question. Xiao’s case rests on supply — cheap energy, fast build-out, four billion potential users — while Al Olama’s rests on application, warning that compute consumption alone decides nothing and that firms leaning on public models may be displaced. Shortening the policy review cycle from a year to six months suggests the government itself is not confident the shape of the technology holds still.

What’s next

Watch the federal agentic AI target, due across half of government sectors within two years of its April launch, and the next six-month review of UAE AI plans. Stargate UAE’s one-gigawatt phase is the nearest measurable test of the delivery claim.