Water, not chips or money, could be the biggest barrier to the Gulf’s AI ambitions

1

The Pitch

Nvidia’s earnings are a positive sign, online chaos not so much

Good morning and welcome to EnterpriseAM AI + Innovation, where we hope you’re looking forward to the weekend as much as we are.

We’re starting today with another split screen.

On one side, there’s the uproar at Nvidia, which seems to be entering a not-worst-case scenario phase. The chipmaker released earnings and forecasts that beat expectations on Tuesday. “Our momentum is real. Our demand is accelerating. Our supply is growing. Our customers and our partners are with us,” interim CEO Colette Kress said on her first results call.

Yes, the company’s stock price is down about 30% since the first revelations that former CEO Jensen Huang was having an affair with an alleged Chinese spy, including a roughly 2.9% slide yesterday, the day after the earnings release.

There are also a lot of shoes yet to drop, including congressional probes focused on Huang and Nvidia and on Commerce Secretary Howard Lutnick and Cantor Fitzgerald, the financial services firm he formerly ran — both men were big advocates of AI chip sales to the Gulf and to China. But the numbers suggest things aren’t remotely as dire as they could have been.

Anthropic’s IPO, which starts trading today, is another green flag. More on that below.

On the other screen, there’s all the online chaos — a global Microsoft Outlook and Teams outage and, in the US, network outages at Chase Bank and on classified NSA networks, and the longest flight grounding since 9/11. Plus a supply-chain cyberattack that planted malicious code on the websites of 178 major organizations. It’s not entirely clear why all these network problems and attacks are happening now, but the rapid pace of technological development during the AI age is almost certainly partly to blame.

That’s a stark warning for the whole planet, especially the MENA region, where cyber defenders have struggled to keep pace with AI-enabled hackers.

In most cases, the outages were simply technical failures during upgrades — but they were driven by outdated systems that aren’t prepared for the rapid pace of patching that’s become necessary since hackers’ ability to discover exploitable bugs in tech systems was supercharged by AI. Here’s a good explainer on X from Sen. Mark Warner (D-Va.), the likely head of the Senate Intelligence Committee when Democrats take control of the chamber in January. –Joe

Now, let’s get to today’s Lede, which digs into the pressure Gulf data centers are putting on the region’s fresh water supply.


BUT FIRST: Would you like to have lunch with us in Dubai on Thursday, 10 December? We’re setting a table in DIFC for a small group of readers and business leaders in the UAE. If you’d like to be there, just hit reply to this email, tell us who you are and what you do, and we’ll take it from there.

Wait, lunch? Yes, lunch. Two hours, one table, no panel, no slides, just an open conversation among peers. We want to hear how the people running businesses in the UAE see the next twelve months: what’s working, what’s getting harder, and where the money is actually going now that the region is open for business again. Expect good food, good company, and the kind of conversation nobody has on stage.

Seats will be strictly limited and the lunch is invitation-only, but we’re planning on doing this more often.

Everything that will be said at the table stays at the table under the Chatham House rule. Drop us a note and we’ll be in touch soon.

2

The Lede

Water, not chips or capital, may be the biggest barrier to the Gulf’s AI ambitions

As the Gulf strives to become a global hub for AI data centers, the region’s limited water supply is probably its biggest barrier. Computing currently consumes just a small portion of our region’s fresh water supply, but that’s sure to change between now and 2030 as the data center building boom expands — both in the GCC and globally.

That’s a challenge for the Gulf, where high temperatures mean data centers need lots of water for cooling — and where fresh water is notoriously scarce. GCC nations all fall below the WHO-defined water scarcity line. And the water they do have is also needed by other sectors, especially agriculture, and for residential needs.

Gulf nations’ fresh water access is also looking especially precarious now. Most countries here rely on desalination plants for 70% or more of their drinking water, and those plants have been targets during the Iran war. (Iranian strikes have hit desalination plants in Bahrain and Kuwait.) That means a single successful drone strike could cut off a significant portion of fresh water supplies, dramatically hampering the ability of the region’s AI industry to meet customer demands.

Facing a wall: Figuring out a durable solution is enormously important to very expensive efforts by the UAE and Saudi Arabia (more than USD 100 bn and counting) to reorient their economies around AI and technology.

“These countries’ push for AI data centres as a pillar of economic diversification will only contribute to the already increasing and unsustainable water demand,” Naser Alsayed, senior fellow at the Dubai-based Fiker Institute think tank and an expert on the Arab Gulf states, tells EnterpriseAM.

Data centers’ reliance on desalinated water is particularly challenging because it adds “pressure on already constrained water resources, creating competition with household consumption and other economic sectors — such as industry, tourism etc.,” Alsayed warns.

The risk is especially high because the GCC’s big pitch to Western partners is that its abundant capital and energy supplies, alongside easy permitting, mean it can get data centers up and running fast. The Gulf’s “competitive advantage will increasingly depend on how sustainably it can power and cool its AI data centers,” and whether the region can sustainably deploy its resources at scale, Jessica Obeid (LinkedIn), energy consultant and founding partner at New Energy Consult, tells EnterpriseAM.

The numbers: Data centers worldwide used around 560 bn liters of water in 2023, the International Energy Agency estimates. That figure is expected to roughly double by 2030 to about 1.2 tn liters annually, the IEA says.

BUT: Only about a quarter of that is fresh water used to cool data centers. About two-thirds is water used to produce the energy that runs data centers. In the Gulf, conventional power plants are mostly cooled with seawater, which isn’t in short supply.

Climate change will exacerbate things, the IEA says, recommending that data centers should be set up in areas of low water stress. In other words, not deserts.

Data for the Gulf specifically is spottier. In the UAE, the Indian market research firm Mordor Intelligence estimates a jump from about 24 bn liters in 2025 to about 61 bn liters by 2030. The firm estimates an even steeper jump in Saudi Arabia, from about 20 bn liters in 2025 to about 88 bn liters by 2030.

How much data center capacity will be built during that period is unclear. The UAE is in the process of building a 5 GW AI campus, likely spread in clusters across the country, led by the national AI champion G42 and US partners — but there is no current timeline for its completion. Saudi AI champion Humain is targeting roughly 3GW by 2030 and 6GW by 2034, though it says it won’t build ahead of demand.

The most significant way data centers use water is to cool servers that run super hot — either to chill air that’s blown over the servers or, increasingly in the Gulf, by running cold, chemically treated water through plates that sit directly on top of computer chips. (We’ll get into more detail in our Building Blocks section, below.)

AND: The GPU chips housed in AI data centers require far more cooling — and thus more water for cooling — than the CPU chips that run traditional data centers.

A lot of fresh water also gets used in the construction of massive data centers — for things like concrete mixing, dust removal, and prepping cooling systems.

The electricity factor: There’s a related problem of electricity capacity. For most of the Gulf, electricity supplies are sufficient but often strained, especially in summer, when everyone’s running the air conditioning at full blast.

That means there will need to be a significant buildout of generation and transmission capacity if Gulf nations build new desalination plants to supply data centers’ fresh water needs. “The challenge lies in whether electricity networks and water infrastructure can expand at the pace of AI investments,” Obeid says.

So, what’s the fix? Most recommendations focus on using water more efficiently, recycling water within the data center, and relying on wastewater rather than the fresh stuff.

“Ultimately, we want data centers’ water usage to exist in a completely closed loop,” Stefan Umiastowski (LinkedIn), SVP and CEO for IMEA at Ecolab, tells EnterpriseAM. “If it’s not addressed in a different way, then we’ll be facing a wall,” he adds. Ecolab sells systems that recycle water within data centers and other production facilities. The ultimate goal, Umiastowski says, is a data center that uses just as much water as a standard washing machine — though that’s a long way off.

Cue the government: Achieving these kinds of efficiencies will likely require a significant push from policymakers, rather than relying just on data center operators themselves, Umiastowski says. Officials in the UAE and Saudi Arabia have routinely cited the importance of fresh water access for economic growth, but have not released formal plans to reduce or manage water consumed by AI data centers.

A late mover advantage? In some ways, the (relatively) new AI buildout in our neck of the woods can work in our favor. Building in water-recycling and -saving technologies in new data centers is significantly easier than retrofitting them into data centers that are already up and running, Umiastowski says.

There’s also a possibility that the Gulf’s strained water supply will force technological developments that lower data centers’ reliance on fresh water. “The region is no stranger to innovation,” Alsayed, at the Fiker Institute, says. He predicts that GCC countries could “leverage the challenges posed by water scarcity to position themselves as hubs for technological innovation in making the sector more sustainable.”

3

The Business of AI

Anthropic’s USD 81 bn IPO starts trading today — and Qatar’s QIA has asked for more shares

Anthropic’s shares start trading on the Nasdaq today, testing how bullish investors are on AI after a turbulent couple weeks.

Here’s what we know so far:

Early signals are good. The company was planning to price its shares at USD 70-78 apiece, but actually priced them at USD 83 last night, the Claude maker said. It also bumped up the number of shares on sale to 980 mn from 825 mn — giving it a valuation just a tiny bit north of USD 1.5 tn, CNBC reports. As of Sunday, there was demand for more than 7x more shares than was on offer, meaning many investors won’t get all the shares they asked for — several investors had asked for USD 10 bn or more worth of shares each, Bloomberg noted earlier this week.

Big spenders: Google and Amazon will own about 28% of the company, per TechCrunch.

Gulf funds are also near the front of the line because Anthropic is offering shares to its existing investors first, Bloomberg says. That includes Abu Dhabi’s MGX and the Qatar Investment Authority, which already own stakes. And the QIA has asked for more: “We did submit and in a couple of days we’ll see how the results are,” CEO Mohammed Al-Sowaidi told Semafor on Wednesday. It’s not yet clear whether MGX is buying more.

It’s now the biggest IPO ever: Anthropic and its co-founders are selling roughly USD 81 bn worth of shares, topping the USD 75 bn SpaceX raised in June. Seven of the company’s compute suppliers, including Google and Nvidia, are also buying USD 26 bn in a separate private placement, taking the total raised to about USD 107 bn.

4

AI in Business

Most companies use AI and few profit from it. The fix, executives say, is to start small.

Companies rolling out AI should set realistic expectations, start with small pilots, and prove the technology works before spending big on it, tech executives advised at the Cairo ICT conference this month.

Why it matters: Just because you use AI doesn’t mean you’ll make (or save) money with it. Nearly nine in ten people McKinsey surveyed say their companies use AI regularly, but only 37% say it’s done anything for profits, the consultancy found. Research firm Gartner predicts more than 40% of agentic AI projects — AI that does tasks on its own — will be scrapped by the end of 2027.

Expectations are the first trap: Companies often build a custom AI tool to do one specific job, then judge it against ChatGPT, a tool that’s been mass-tested for years, says Mohamed Ashraf (LinkedIn), who heads AI and business software development at Zone. Unsurprisingly, the first reaction is often disappointment. But that usually isn’t the AI’s fault, because the company often picked the wrong task to turn over to AI in the first place, Ashraf says.

Start small: GPUs, the powerful chips that AI runs on, don’t come cheap. So if you ask your company for a big budget to buy them upfront, the money folks “will remain skeptical,” says Dr. Hatem Elattar (LinkedIn), head of industrial AI at Elsewedy Electric. His advice is to rent computing power from a regional cloud provider for a small pilot and show results first. “Once they see tangible value delivering an operational impact, the business will be ready to invest,” he says.

Other advice:

  • Keep the sensitive stuff at home: Elattar favors a hybrid setup. Data that’s especially sensitive stays on a company’s own servers, while everything else can run in the cloud.
  • Use the front door: AI shouldn’t act “like the company owner’s privileged relative who enters unchecked, claims full access, and demands visibility over every database,” Ashraf says. It should follow the same access rules as any other system, he says. Companies should also start off with AI only being able to read data, not change anything, he says.
5

Building Blocks

Direct-to-chip water cooling is becoming the standard for AI data centers in the Gulf’s heat

Artificial intelligence is the Gulf’s hottest investment right now in more ways than one: How do you keep a data center cool in one of the world’s hottest regions?

Let’s start with the basics. AI data centers can draw 100 MW of power or more — two to four times what a traditional data center might burn. That produces a lot of heat — heat that needs to be dissipated with air or water.

A decent ballpark figure is that cooling comprises about 20-40% of the energy use and about the same percentage of operating cost of a large data center. Those figures are higher in less efficient data centers and lower in more efficient ones.

Cooling is even more critical in hot regions like the Gulf, where temperatures don’t take a break at night. “In cooler climates, you have what I’d describe as free cooling, especially at night,” Maria Goodpaster (LinkedIn), a partner at the consulting firm McKinsey & Company who advises companies on AI infrastructure, tells EnterpriseAM. “In hot climates like the Gulf, you just don’t have that as much. You have to use energy for most of your cooling.”

From air to water: Traditionally data centers were cooled with air cooling systems, similar to big air conditioners, but those can’t keep up with the Gulf’s cooling needs. There are various methods of water cooling, but the most common in large AI data centers is direct-to-chip cooling. That means a metal plate with channels carved inside of it is placed directly onto the chip, then cold, chemically treated water runs through those channels, keeping the chip cool.

A more experimental method is called immersion cooling. “This is like putting the server components directly into a big tub of special, non-conductive fluid [resembling] mineral oil,” Shaolei Ren (LinkedIn), a professor of electrical and computer engineering at the University of California, Riverside, tells EnterpriseAM. Theoretically, immersion cooling will be more efficient in many cases (if it’s widely adopted), but it also creates problems of its own. For example, imagine trying to repair a malfunctioning server rack right after it emerges from a big tub of goo.

Direct-to-chip cooling is likely to be the industry standard for big AI data centers for at least the next decade, Goodpaster says.

These systems minimize water waste by operating as much as possible on a “closed loop.” That means the same chemically treated water flows across the chips, sucks out the heat, then gets cycled out of the server racks and cooled before it’s cycled back in again.

But: Separate water is also often used in the process of cooling down the chemically treated water before it repeats its journey. And those facilities still consume water for other purposes, such as cleaning and plumbing. Ironically, older, air-cooled systems often waste more water than water-cooled systems because they deliberately evaporate water that’s used to cool the air that goes on to cool the servers.

Best-case scenario: Microsoft has boasted that its new data center design, launched in 2024, “consumes zero water for cooling,” reducing water consumption by 125 mn liters per data center per year, the company said.

6

The Security Beat

Data centers are critical national infrastructure, UAE cyber chief warns

Data centers are “critical national infrastructure, and they have to be secured as sovereign assets from the first design drawing,” Dr. Mohamed Al Kuwaiti, leader of the UAE’s Cyber Security Council, said at the opening of the Data Center World Middle East conference in Abu Dhabi this week. He also warned that “protection has to reach past the data hall to the substations, cable routes and subsea links that keep a facility running, and to every vendor with remote access to its systems.”

The comments came as the UAE is investing USD tens of bns in growing its data center capacity while simultaneously facing the threat of physical attack from Iran and — like everyone else these days — being pummeled by waves of AI-enabled cyber threats.

“The UAE aims to be the world’s most trusted digital hub. Operators, technology providers and government earn that trust by working together,” Al Kuwaiti said.

7

Deal Tracker

Etisalat Business plans to rent out GPU computing power to Egyptian companies

Etisalat Business plans to launch a GPU-as-a-Service (GPUaaS) offering in Egypt, which would allow renting out cloud computing power to local companies and developers under a three-way MoU with investment group M Empire and the Startup Egypt foundation, the country’s cabinet said in a statement. That matters because the specialized GPUs needed for a lot of AI work pose a steep upfront cost for Egyptian startups and enterprises, which often end up renting capacity abroad at an added cost.

Uhm, Enterprise? Isn’t that a … data center? You’d think so, but no. A data center is essentially the building. A developer puts up the structure, cooling, connectivity, security, and power. Someone else typically buys the GPUs, puts them in racks, and rents out the processing power. That could be the data center’s backer, but generally isn’t. A GPUaaS (can we all agree we don’t want to use that acronym again?) provider is generally someone who already has access to a data center, then buys the chips and rents out processing power by the hour or the month.

The draw: Pricing GPUs in EGP can be far cheaper for Egyptian companies, sparing them the cost of accessing hard currency, paying conversion fees, and dealing with swings in the exchange rate.

A smaller Egyptian startup is already renting out GPUs: Chipverse has closed an undisclosed pre-seed round from regional angel investors and tech entrepreneurs, Technotime reports. Founded in 2024, the startup rents out servers and GPU clusters hosted locally and priced in EGP. It plans to put the new capital toward doubling its GPU capacity, launching an API layer that lets developers plug directly into its infrastructure, and rolling out creative tools for content creators and production studios, the company says.

AND- At the other end of the scale: SpaceX is raising USD 40 bn in debt to buy Nvidia chips for its AI business, Bloomberg reports. The debt won’t sit on SpaceX’s own books. A separate company backed by investment firm Apollo will buy the chips and sell bonds to pay for them. Part of that debt could go on sale as soon as this week.

8

The Global Beat

Microsoft unveils six AI models of its own as it tries to rely less on OpenAI

Is Microsoft ready to stop relying on OpenAI? It’s getting closer. On Tuesday, Microsoft unveiled six of its own AI models that it claims compete with top models from OpenAI and Anthropic at a fraction of the price, Axios reports. Its new coding model, for example, is almost as good as Anthropic’s wildly popular Claude Opus 5.5, but costs about 20 times less, Microsoft says.

The caveat: Microsoft picked the tests behind those claims.

Watch the trend: Microsoft wants to make more of its own AI instead of buying it from other companies. Besides these six smaller, targeted models, Microsoft is also working on a bigger, all-round model, AI chief Mustafa Suleyman says. Last month, Microsoft said it plans to deploy potentially more than a million of its own AI chips next year, to cut costs and reliance on Nvidia.

9

The Chaser

OpenAI’s new GPT-6.2 models get a rough reception from AI power users

Mixed reviews: OpenAI launched two new models on Tuesday: GPT-6.2 Astra, its most powerful model, and the cheaper GPT-6.2 Sol. The releases come about seven weeks after the company scrapped the release of GPT-6.1 Astra over safety concerns.

Early reviews from AI power users are rough, VentureBeat reports. Developer Theo Browne called Astra “the worst model OpenAI has ever released.” Investor Matt Shumer said the models were “essentially unusable.” OpenAI’s own tests, for what it’s worth, show Astra beating its predecessor across the board. Sam Altman wrote on X that “we have more work ahead of us on making the models nice to work with.”

Cleared for takeoff: Remember Google buying defunct Spirit Airlines’ old emails and spreadsheets to train its AI? A US bankruptcy judge approved the USD 10 mn deal on Tuesday, Reuters reports, after Google agreed to extra privacy protections.

There are no films, to our knowledge, that deal directly with selling off a defunct airline’s data, but there are plenty about old-fashioned corporate drama like Wall Street, Other People’s Money and, best of class on the corporate espionage side, Michael Clayton — IYKYK: “I’m Shiva, the god of death.”

That’s it for today! Like what you’re reading? Feel free to invite some friends. Or drop us a line at [email protected]. We’ll be back in your inbox on Tuesday.


November 2026

17-19 November (Tuesday-Thursday): Bengaluru Tech Summit brings businesses, startups, researchers, and policymakers to Bengaluru, India, under the theme “AI & Beyond.”

17-20 November (Tuesday-Friday): Microsoft Ignite runs in San Francisco and online, covering Microsoft’s AI, cloud, security, and data products.

30 November-4 December (Monday-Friday): AWS re:Invent brings Amazon’s cloud and AI conference to Las Vegas, with product announcements and sessions on deployment, costs, and security.

December 2026

6-13 December (Sunday-Sunday): NeurIPS holds its main AI research conference across three simultaneous sites: Sydney (6-12 December), Atlanta and Paris (9-13 December). It offers a look at advances that could shape the next generation of commercial AI.

7 December (Monday): GITEX Summit, Dubai World Trade Centre, Dubai.

8-11 December (Tuesday-Friday): GITEX Global and Expand North Star, Dubai Exhibition Centre, Expo City Dubai, Dubai.

January 2027

6-9 January (Wednesday-Saturday): CES returns to Las Vegas, where consumer electronics makers and chipmakers show off AI devices, PCs, and connected cars.

18-22 January (Monday-Friday): World Economic Forum Annual Meeting brings political and business leaders to Davos, Switzerland, where AI policy and investment are regularly on the agenda.

31 January-3 February (Sunday-Wednesday): Web Summit Qatar holds its fourth edition at the Doha Exhibition and Convention Center, bringing together startups, investors, and tech companies.

February 2027

1-3 February (Monday-Wednesday): World Governments Summit gathers heads of state, ministers, and tech executives in Dubai to talk about government and the future of technology, including AI.

16-23 February (Tuesday-Tuesday): AAAI-27, the Association for the Advancement of Artificial Intelligence’s annual research conference, takes place in Montréal, Canada.

March 2027

23-24 March (Tuesday-Wednesday): Global AI Show opens at Space42 Arena in Abu Dhabi, with enterprise AI, computing capacity, and sovereign AI on the agenda.

Now Playing
Now Playing
00:00
00:00