Gulf’s AI strategy is built on more than sand
The author is a Reuters Breakingviews columnist. The opinions expressed are her own.
By Karen Kwok
ABU DHABI, Nov 13 (Reuters Breakingviews) -When it comes to major tech investment, Gulf states have a chequered record. The United Arab Emirates and Saudi Arabia, after all, sank $60 billion into Masayoshi Son’s first Vision Fund, for decidedly slim returns. But with artificial intelligence – where losing stacks of money is a very real risk for investors globally – Abu Dhabi in particular looks to have the makings of a valid strategy.
AI is still in its infancy. But it’s reasonable to assume the globe’s two biggest economies, China and the United States, will also be AI superpowers. Take OpenAI, the $157 billion startup which builds large language models and the conversational tool ChatGPT; or Nvidia NVDA.O, which designs graphics processing chips crucial to power data centres and AI models. AI startups in the United States have raised $70 billion from investors so far this year, according to PitchBook, while the Chinese had raised $6.5 billion.
The Gulf lacks that sort of private sector heft for AI: VCs only invested $700 million in AI startups in the Middle East so far this year, PitchBook says. But it does have lots of state oil money. This year the UAE announced a $100 billion fund, MGX, comprising $30 billion of equity provided by U.S. groups BlackRock BLK.N and Microsoft MSFT.O as well as Abu Dhabi sources. MGX is effectively a version of the UAE’s more established $300 billion sovereign wealth fund Mubadala, but focusing solely on speedily investing in AI companies around the world like OpenAI. Saudi Arabia is also planning a similar sized fund, and has hooked up with U.S. Big Tech giant Alphabet GOOGL.O to build an AI hub.
The UAE government, which consultants view as two years ahead of Saudi on AI, sees MGX as just one prong of a comprehensive AI strategy. Another key component is G42, a 25,000-strong company Abu Dhabi has used to build data centres and cloud facilities in the region and in Africa. G42 is also developing AI applications in the healthcare, space and data sectors via offshoots like M42, Space42 and Core42.
Arguably the most important piece of the jigsaw, however, is academia. Researchers at the government-funded Technology Innovation Institute represent the bit of “Abu Dhabi AI Inc” where the emirate developed Falcon, which along with France’s Mistral and Meta Platforms’ META.O Llama is one of the world’s most advanced open-source LLMs. TII, consultants say, is what differentiates Abu Dhabi from merely being a pile of cash and hype, and being a meaningful AI player.
All the Gulf’s nascent AI plans have the same problem, though. While Mubadala has an 82% stake in $24 billion chipmaker GlobalFoundries GFS.O, that entity isn’t supplying the vital Nvidia-like advanced chips that can be used with data centres to train the most cutting-edge LLMs. And for all their low taxes and sunshine, the Gulf states can’t just create a Silicon Valley-style ecosystem of innovative and competitive AI experts from scratch.
Abu Dhabi and Saudi’s strategies for dealing with these headaches take in the fact they have leverage of their own. Hyperscalers like Amazon AMZN.O, Alphabet’s Google and Microsoft may have interests in LLM providers like OpenAI and Anthropic, but their immediate business opportunity is to entice the likes of G42 to plug into their data centres to build, test, run and power their AI models. The U.S. government, meanwhile, wants nascent Gulf hubs like Abu Dhabi to not get as close to China, its main AI rival. That explains why Microsoft this year invested in G42 and MGX, and why the UAE is opening up its data centre market to U.S. hyperscalers who are prepared to come to the region. The implicit quid pro quo is that the UAE will sit firmly in the U.S. AI camp rather than China’s, and that the U.S. will make it easy for the UAE to access Nvidia chips.
Still, even if Abu Dhabi can get all the chips it needs, it has to decide what its AI capabilities are for. On this front, the emirate appears to have a plan. According to AI experts Breakingviews spoke to in Riyadh and Abu Dhabi, Abu Dhabi’s leaders recognise that chasing the next big LLM models in rivalry with Microsoft-backed OpenAI or Amazon-backed Anthropic risks incinerating huge amounts of money. The Gulf may have the same tens of billions of dollars that the hyperscalers have to spend on data centre capex and power, but all these players are running the risk that their LLM doesn’t gain traction, leaving them with miserable returns on their capital.
That’s why the UAE is focusing on a less ambitious path. More than half of web content is in English, while Arabic text makes up only 1% of the content on the World Wide Web. The shortage of high-quality Arabic language content available gives the Gulf a reason to build their own database and AI models. That’s why researchers at G42’s Inception arm have been trekking to libraries around the region and literally scanning in pages of Arabic text to put together their own model, JAIS.
The logic is simple: training an English LLM and then translating it into Arabic costs a third more than training one using solely Arabic inputs, according to companies which develop AI solutions in the region. It also removes the sort of Western bias inherent in most AI models. Meanwhile, Saudi Aramco claims it can deliver power domestically around 13% cheaper than the average 7.8 cents per kilowatt hour paid by data centres in the United States. The upshot is that Gulf AI players may be able to train their models more cost-efficiently, raising scope for the 400 million-strong Arabic-speaking market to buy their products as and when the AI sector as a whole starts to make money.
The other part of Abu Dhabi’s strategy should yield more immediate dividends. Rather than chasing the Google Search or ChatGPT “business to consumer” market, the emirate is more interested in making its businesses more efficient. AIQ, a joint venture of Abu Dhabi oil giant ADNOC, G42 and Presight, has built specified AI software called “RoboWell” to improve the productivity of its gas wells through real-time autonomous adjustments, an undertaking that has already raised production volumes by 5% and driven cost savings equivalent to a 30% bump to operating profit margins. Malaysia’s Petronas has signed up for AIQ’s range of technologies. G42 and Mubadala’s health business M42, meanwhile, has built a medical AI model that achieved a 95% score in a U.S. medical exam, and built an AI solution to screen early signs of disease which helps lower costs for radiologists by 20%.
The sensitive geopolitical nature of the AI race, and the Gulf’s ongoing need for chips and knowhow, means both the UAE and Saudi may struggle to become global AI leaders. But the UAE in particular seems to be making logical decisions in order to become a meaningful player. If it sticks to the more sensible niches, AI could be one tech investment that isn’t built on sand.
Follow @karenkkwok on X
Graphic: VC funding of Middle Eastern AI firms remains tiny https://reut.rs/4en1AVL
Graphic: Government investments in AI companies https://reut.rs/4fHu3XE
Editing by George Hay and Oliver Taslic
Related Assets
Latest News
Disclaimer: The XM Group entities provide execution-only service and access to our Online Trading Facility, permitting a person to view and/or use the content available on or via the website, is not intended to change or expand on this, nor does it change or expand on this. Such access and use are always subject to: (i) Terms and Conditions; (ii) Risk Warnings; and (iii) Full Disclaimer. Such content is therefore provided as no more than general information. Particularly, please be aware that the contents of our Online Trading Facility are neither a solicitation, nor an offer to enter any transactions on the financial markets. Trading on any financial market involves a significant level of risk to your capital.
All material published on our Online Trading Facility is intended for educational/informational purposes only, and does not contain – nor should it be considered as containing – financial, investment tax or trading advice and recommendations; or a record of our trading prices; or an offer of, or solicitation for, a transaction in any financial instruments; or unsolicited financial promotions to you.
Any third-party content, as well as content prepared by XM, such as: opinions, news, research, analyses, prices and other information or links to third-party sites contained on this website are provided on an “as-is” basis, as general market commentary, and do not constitute investment advice. To the extent that any content is construed as investment research, you must note and accept that the content was not intended to and has not been prepared in accordance with legal requirements designed to promote the independence of investment research and as such, it would be considered as marketing communication under the relevant laws and regulations. Please ensure that you have read and understood our Notification on Non-Independent Investment. Research and Risk Warning concerning the foregoing information, which can be accessed here.
