Happy THURSDAY, wonderful people. It’s been a brisk week here at EnterpriseAM World Headquarters, and we couldn’t be happier to be on the cusp of the weekend. So much the better that it’s nearly Halloween — happy trick-or-treating to all those who observe (or whose kiddos will drag them out).
I hope to spend the weekend unwinding, but work beckons, so I know I’ll be thinking and reading and writing about business and AI even as I try to finish up Andrew Ross Sorkin’s 1929 (highly recommended) or distract myself with the new season of The Diplomat on Netflix (apparently the best yet).
What’s gnawing at my brain: To write is to think — and I worry that too many of us in the business community are not-so-slowly losing our ability to do both, at least on the job.
One CEO I know recently kicked off his weekly exco meeting with a little bit of performance art, reciting poetry of a form to his company’s leadership:
The gate is the loop run clean to a defensible recommendation — not the signature. The rock does not ride on a candidate's pen.
One date means bulk-set close dates
Meetings and proposals go on the weekly scorecard, not in the gate
A hypothesis wearing a rule’s clothing
Instrumentation is the unlock.
inside it
a dial worth turning
An idea that earns its place.
A plan built. Honest framing. A named responsibility.
Load bearing.
^^ All of it stuff he had pulled out of his team’s weekly reports.
Prose from an LLM can be just … exhausting to read. You get to the end of a sentence and while you know what each and every word means, you have no idea what the sentence itself is actually trying to say. Read it again and you might grasp some spectral outline of what’s intended, but actual meaning? Not so much.
Your Slack / Notion / inbox / Asana / whatever are all overflowing with Claudeslop written in Claudish, and your new job is trying to decipher what it actually means.
The worst thing you can do is point your own robot at it and ask it for help. You lost a significant chunk of meaning as an LLM went to work on your trusted lieutenant’s prompt and data inputs. Handing the output off to your robot and asking for a summary will just dilute meaning even further. (And God help you if there’s an Arabic name or 10 in the conversation from a meeting transcript.)
Ever wonder why robots write that way? Three scholars recently set out to answer just that, and their findings are in a preprint released last week. They ran an analysis of output from Opus, Fable, and Sol — nine LLMs in all — and found our robot helpers love to pair animate verbs with inanimate objects, stack nouns together, you name it. If you, too, are being driven ‘round the bend, add “Why is AI-generated language so hard to understand?” to your weekend TBR pile.
I hope that “lands” for you — because the torrent of Claudeslop (and Chatslop and Geminislop) now flowing through corporate life is more insidious than bad writing. It also suggests a deficit in thinking.
I’m starting to worry that it’s a fallacy that (most) humans want to be in the loop. Because if we do? Well, we sure as hell haven’t done a very good job of it in these early days of the AI revolution.
The danger of Claudeslop isn’t just that you’re getting machine language — it’s that you’re also getting unquestioned robot judgement masquerading as that of your team, boss, whoever.
That’s not just me, an angry middle-aged guy yelling at clouds:Research out this week from Anthropic found that in a controlled experiment of 50 analysts asked to give business advice on the same case, “77% of analysts used the agents’ own arguments — sometimes word for word — but only 15% credited the agent as the source of those arguments.”
In other words: They pretty much adopted the agents’ thinking wholesale — and may not even have been aware of it.
That’s it for me. I’m handing the pen back to Joe next week. And speaking of Joe: You won’t want to miss this morning’s Lede, wherein he writes about what it might mean for the Gulf’s AI ambitions should Democrats recapture the House and Senate when Americans head to the polls on 3 November. Have a wonderful weekend, everyone. –Patrick
A series of Trump administration deals that boosted the Gulf’s outsized AI ambitions is still roiling US Democrats as midterm elections approach on Tuesday, 3 November, signaling trouble could be afoot if the party seizes power in Washington.
Democrats can’t kill the deals, which gave UAE AI champion G42 and its cloud division Core42 access to the most advanced artificial intelligence chips without an export license and approved up to 35K of those chips for Saudi Arabia’s AI champion Humain.
How real is the threat? Republicans are in especially dire shape as the elections approach, with most polls suggesting Democrats will easily win control of the US House and have a good chance of also seizing control of the Senate.
But they’rethreateninginvestigationsif they win control of Congress — focused on accusations the deals could leak AI secrets to China and that they don’t sufficiently protect against using AI for surveillance.
Quid pro quo? They also claim the UAE agreement may have been a quid pro quo for a USD 500 mn investment in World Liberty Financial, a Trump family cryptocurrency venture, by the firm Aryam Investment, which is backed by UAE National Security Adviser Sheikh Tahnoon bin Zayed Al Nahyan. Tahnoon, a member of the UAE royal family, is also chairman of the boards of both G42 and MGX, a sovereign-backed investment fund that focuses extensively on AI.
If a Democrat wins the presidency in 2028, the party could clamp down on future chip deals with Gulf allies — either by refusing to green-light new sales of the highest-powered Nvidia chips or limiting access to the most powerful and up-to-date AI models.
Restricting chip sales could be complicated by existing agreements, though. The UAE’s access to top-shelf silicon without an export license expires in April, before Trump leaves office. If the deal is extended, it would be extremely difficult to roll it back before a future expiration date. G42 and Core42 must become US companies by 6 April 2027 or seek fresh authorization to retain license-free chip access.
Any move to limit Gulf chip access would be a significant blow to the region as it aims to shift from oil to computing power as its major export commodity. The consultancy PwC has estimated that AI could add USD 135 bn to Saudi Arabia’s economy and USD 96 bn to the UAE’s by 2030, or roughly 12-14% of GDP. Humain declined to reveal specific requirements of its deal with Nvidia or to say how many chips had been delivered so far. G42 did not respond to an EnterpriseAM request for comment.
Pushing for guardrails: “There is a potential model for US technology collaboration with Gulf partners — but it has to be predicated on technology security, strong protections against misuses of US-origin technology for surveillance or repression, and tangible benefits to the United States. Instead, the Trump administration has predicated these engagements on transactionalism,” Sen. Mark Warner, the top Democrat on the Senate Intelligence Committee, tells EnterpriseAM.
Details:
Sen. Elizabeth Warren, the top Democrat on that chamber’s banking committee, blasted the US Commerce Department’s decision to give G42 license-free access to advanced chips, accusing the administration of ignoring “concerns about the diversion of sensitive technology to China and other national security risks.”
Rep. Ro Khanna, a California Democrat, has pledged to investigate that deal as a top member of the House Oversight Committee and a special House committee overseeing US-China policy.
Top Democrats on the House Foreign Affairs, Intelligence, Armed Services and Oversight committees have urged the Commerce Department’s internal watchdog to probe the Trump family’s financial interests in the agreement.
“Decisions about exports of advanced American AI chips should solely be based on economic competitiveness and national security considerations, and they should certainly never be made to make money for your own business,” Rep. Sydney Kamlager-Dove, the top Democrat on the House Foreign Affairs Committee panel with jurisdiction over South and Central Asia, tells EnterpriseAM.
If Democrats win control of Congress, they might also cooperate on new AI export controls with Republican security hawks, who are concerned about US AI technology leaking to China. “[Trump] probably would not listen to the Democratic leadership, but … he would pay attention to the Republican leadership telling him they're going to support the Democrats on something like this,” William Reinsch (LinkedIn), a former US Commerce Department official who was in charge of export control policy during the Clinton administration, tells EnterpriseAM.
But the Democratic pushback also comes amid a broader intertwining of financial and AI interests that will make any US decoupling with the Gulf highly complicated. Gulf sovereign wealth funds have invested billions in US AI companies, including OpenAI, Anthropic and xAI. More broadly, the president picked the Gulf for the first major international trip of his second term, where he announced USD 200 bn in new US-UAE business deals.
SOUND SMART- Midterms are the US elections held halfway through a president’s term — putting the entire 435-seat House of Representatives and roughly a third of the Senate up for grabs. They double as a verdict on the incumbent (whose party almost always loses seats) and can hand the opposition control of Congress — throttling or accelerating a president’s agenda for the last two years of his term.
The Trump Administration has taken a mostly hands-off approach to AI, urging the major labs to police themselves rather than providing government oversight. His administration has argued that the United States has a better shot at staying ahead in the AI race if it makes the rest of the world dependent on US technology than if it hoards that technology for itself – making Chinese AI more attractive in the process. “I’m not going to stifle growth of something that will be bigger than the industrial revolution,” Trump said during a September speech at the United Nations.
But Democrats and many Republicans argue that this laissez-faire approach could endanger the US — by allowing high-powered AI technology out into the world before it’s sufficiently vetted and by making it easier for China to copy US models.
To be sure, the Saudi and UAE deals don’t ignore security concerns. They include “rigorous security and reporting requirements” aimed, in part, at preventing AI technology from ending up in Chinese hands, according to the Commerce Department, which approved the sales. But the risks of “leakage” to China are inevitably high given the long history of tech partnerships between Saudi Arabia, the UAE, and China — and the Asian giant’s long, documented history of government-backed digital theft.
Decoupling: UAE officials have said they’re “decoupling” from Chinese tech in their AI infrastructure as a condition of AI cooperation with the United States, but the nations remain tightly linked in other areas of technology. That decoupling has included G42 selling off all its investments in Chinese firms. Saudi Arabia has not made public decoupling pledges.
But narrowing the terms of the current chip deal or restricting future ones would create a whole separate set of problems. For starters, even if Democrats sweep the midterms, they couldn’t make major changes until a Democrat is in the White House more than two years from now. By that point, US and Gulf tech firms will be even more interlinked and American firms may be more dependent on Gulf capital.
“We’re talking about USD bns in investments, joint data centers and deep partnerships between American tech giants and Gulf sovereign entities. A future administration would face heavy industry pushback,” Mohammed Soliman (LinkedIn), a senior fellow focused on technology at the Middle East Institute, tells EnterpriseAM.
That means future administrations are more likely to tinker around the edges of chip deals — ramping up security protections and access restrictions — or to restrict access to the latest AI versions rather than cutting off cooperation entirely.
Devil in the details: “The bigger question is not whether controls remain in place, but how aggressively a future administration applies them to third countries and what safeguards it requires for AI-related partnerships,” Leah Siskind (LinkedIn), an AI research fellow at the Foundation for Defense of Democracies, tells EnterpriseAM. “Any Democratic administration is likely to scrutinize technology transfers, cloud access, and chip deployments to ensure they do not create indirect pathways for China to access advanced AI capabilities,” she says.
Plus any moves that hurt the US AI industry are likely to have major repercussions for the broader American economy, which has been mostly stagnant except for AI spending. “AI is such a linchpin for GDP, it’s tough to see putting that at risk,” Paul Meeks (LinkedIn), head of technology research at Freedom Capital Markets, tells EnterpriseAM.
There are also broader geopolitical considerations. Most importantly, Democrats may be wary of punishing Gulf allies after they’ve borne the brunt of Iranian missile and drone attacks during the war with Iran and as they continue to provide a strategic buffer against Iran in the region. The Trump administration announcement about removing license requirements for UAE chip exports specifically cited the nation’s “status as a US major defense partner and its support in advancing US national security interests, including Operation Epic Fury.”
Paying a debt: “Arguably, the severe damage imposed on the Gulf states by this war means that the US is under an obligation to be a supportive ally in the coming years, including facilitating chip exports,” Justin Alexander (LinkedIn), a Gulf business analyst and Director of Khalij Economics, tells EnterpriseAM.
And the damage is far from over: Trump has signaled he intends to renew US strikes on Iran after the midterms, despite the Pentagon’s dwindling stock of missiles. Saudi Arabia is already suffering from a surge in brutal attacks by the Iran-aligned Houthis targeting airports and energy infrastructure, prompting Washington to order non-emergency diplomats out of the kingdom earlier this month.
Finally, the more time passes, the less leverage Washington will have over global AI infrastructure. While US companies are ahead of China in the AI race right now, most experts expect that picture to change — if not by the time the administration takes office in 2029 then quite possibly by 2031 or 2032 when the tranche of Nvidia chips now in service reaches the end of their expected lifecycle. “The AI race is squarely in our favor right now, but the Chinese are super aggressive,” Meeks says. “Down the road, we can expect Huawei to have a chip every bit as good as Nvidia.”
Governments from Abu Dhabi and Riyadh to Cairo are increasingly worried about control of their countries’ data as more and more corporate and state services disappear into the cloud.
But having the servers on your soil only gets you so far, as the war in Iran has shown. Gulf AI projects still depend heavily on American chips and software, as we note in this morning’s Lede (above), and keeping services running through a war adds another wrinkle — somewhere safe to which you can “fail back” if a data center is taken offline by the conflict, as happened to Amazon Web Services earlier this year in the UAE and Bahrain.
Governments and regulated industries can’t move as freely as many private-sector companies. Banking, healthcare and public-service workloads need arrangements that preserve security and control, Rihla Research and Advisory CEO Jesse Marks tells EnterpriseAM Infrastructure. A financial institution just doesn’t have the same flexibility to spin up a server in France that, say, an HR consultancy would.
Data sovereignty rules in Egypt, the UAE, and Saudi risk stifling innovation. Look no further than Egyptian banks, which are having to work with open source LLMs on local hardware using tiny clusters of imported GPUs rather than global-quality data centers abroad.
One workaround now gaining traction is a “digital embassy” — agreements that would allow specific data centers in a specific country to process data under the home country’s legal authority.
No Gulf state has one yet, Marks says. He expects simpler emergency agreements with trusted countries could come first. Some have suggested starting the process with cloud continuity drills and shared data classification rules along with “trusted backup capacity in the United States, India, Singapore, Turkey, and selected European markets,” says Yasir Atalan, deputy director of the CSIS Futures Lab.
A handful of reports out this week are giving us a better picture of how businesses across MENA+ are using AI in offices, shops, and on factory floors. Editing documents leads work-related ChatGPT use in the UAE and Saudi Arabia, while visual design is the number one “use case” (as the software nerds refer to it) in Egypt.
In the UAE, editing accounts for almost one in four work-related messages, according to country data from 3Q published this week by OpenAI. Our analysis of the figures puts it at 24.1%, followed by visual design (6.2%), and interpreting language, cultural, or religious information (5.8%).
Saudi Arabia follows the same order. Editing leads at 15.3%, ahead of graphic design (8.6%), and interpretation (7.5%).
Egypt’s biggest use case is visual design, at 10.6%. Editing comes a close second at 10.2%, followed by developing marketing or promotional materials (7.6%).
The US and UK look more like the Gulf. Editing ranks first in both, at 17.1% and 18.8%, respectively. Visual design comes second — 7.6% in the US and 6.6% in the UK — followed by marketing materials at 5.0% in each. (The OpenAI data excludes messages sent to Codex, its developer-forward platform.)
MEANWHILE- Small businesses in the UAE are easing gently into the water. Emirati expense-management platform Pemo’s first SME Spend Index found that 29% of its active business customers paid for AI in the third quarter of this year. Among companies that were using it last year, AI spending rose 328% year on year and the share of companies paying for it climbed to 35% from 21%.
Most small businesses are not big spenders. The median business in the data paying for AI is coughing up AED 268 a month.
Tech and marketing companies are more AI-forward: Half of technology companies in Pemo’s sample paid for it, compared with 46% of marketing agencies and 17% of construction businesses. (Pemo’s data doesn’t include use of free tools, personal subscriptions, or AI bundled into Microsoft or Google workplace suites.)
AND- People working in the GCC are happy to use AI, but they’re less certain they want a robot deciding whether or not they get a new job. The latest edition of the EXE and Ipsos GCC People Pulse found 83% of respondents in the UAE, Saudi Arabia, Kuwait, and Qatar used employer-provided AI at least occasionally. More than half — 54% — use it daily or several times a week.
We’re actually surprised the numbers are this high: The enthusiasm of the 700 people surveyed cools when it comes to putting their careers in the hands of an AI agent. Some 59% would take part in an AI screening interview, while just 44% were comfortable with AI making the final decision on their job application.
Want an AI agent without having to build one? Dubai Chambers has opened a shop. Its recently launched Chamber AI Hub lets member companies pick from 22 ready-made agents designed to handle jobs including contract review, matching invoices to purchase orders, screening CVs, and drafting business proposals. The chamber says the bots are fluent in English and Arabic and that company data is hosted in the UAE.
Adnoc is doubling down on its bid to make better use of its data using AI: Adnoc Gas is scaling its industrial data platform, while Adnoc Upstream is pooling operating data to improve production, AI vendor Cognite said this week.
BACKGROUND:As we reported earlier this month, oil and gas players are infamous for what one might call a “sub-par approach to data management,” with field and operational data strewn in a variety of formats across devices and systems that don’t talk to each other. AIQ, ADNOC’s joint venture with Presight, told us its engineers spend an average of 100 days on-site consolidating customers’ data and training AI systems — groundwork needed before the technology can deliver on its promise.
What to watch for next: AIQ’s planned unveiling of its Genesis platform at next month’s Adipec conference. The company told us it was testing the product on Adnoc systems, with the aim of automating data extraction, analysis, and visualization across an entire company.
YouTube creators across MENA might be able to get their hands on some affiliate revenue a bit faster after the Google unit launched its Shopping affiliate programme in Saudi Arabia, the UAE, and Egypt. Amazon is on board as its first partner. The program allows creators with at least 500 subscribers to earn commissions on products tagged in videos, Shorts and livestreams. Google uses AI to figure out what’s on screen and suggest affiliate links.
Has China’s AI bubble already burst? Shares in China’s biggest AI and chip companies have lost 40% since July, wiping out about USD 480 bn in shareholder value, Reuters reports, calling the crash a bad sign for already shaky global markets as Anthropic readies a blockbuster IPO.
What happened? Prices got too high, traders say, and investors worry that export bans could block Chinese equipment from offshore data centers, further curbing the industry’s growth prospects.
Signs of a bubble in Western markets are less clear even as worries mount about circular finance and the sheer amount of money that AI companies are chasing, including for Anthropic’s IPO and a massive USD 95 bn equity sale program Meta announced this week.
The latest overnight:
Shares of Alphabetfell after Google’s parent company reported sales and profit figures that fell short of market expectations, even as it more than doubled revenues from its cloud business in 3Q.
Microsoft’s cloud business is stillgrowing. Revenues from the sale of Azure were up 48% and Microsoft Cloud revenues were up 30%.
Samsung’s chip unitmissedprofit estimates, but the company will still be ramping up spending as the AI infrastructure buildout continues.
Apple launched a touchscreen MacBook Pro yesterday. The machine is lightly redesigned and about half a pound lighter and comes with your choice of M5 Pro or M5 Max chips with up to 128GB of RAM and as much as 16TB of storage. They will come in both 14-inch and 16-inch sizes with up to 24 hours of battery life.
You can check out the “MacBook Pro with PROLED XDR Display” (we kid you not — that’s the name) here.
Anthropic’s answer to Dots and Muse is out, but good luck getting one as your very own personal agent if you’re in MENA+.Claude Duet checks connected apps, prepares work, and carries on while you’re away from your machine, asking before it sends, spends, signs, or deletes. It’s a US-only research preview for folks on Pro or Max plans, with a waitlist.
Muse is also “gated” (as the LLMs like to say), though it has expanded beyond the US to Canada and Japan. Dots are more widely available, and OpenAI’s Tibo Sottiaux said on 18 October that the company was working to keep Dots created elsewhere running while their users travel in the EU. The fanbois on X are speculating that Dots should start rolling out in Europe in the coming weeks.
FINALLY- Claude Code engineers Thariq Shihipar and Anthony Morris left Anthropic last weekend. Shihipar was among the product’s most visible builders, and Morris created its desktop version. Neither has said where they’re heading next.
29 October (Thursday): Amazon reports 3Q earnings, including AWS and Bedrock AI growth.
30 October (Friday): UAE businesses with annual revenue of AED 50 mn or more must have an accredited e-invoicing provider in place by this date.
November 2026
7-10 November (Saturday-Tuesday): Cairo ICT runs in New Cairo alongside AIDC Expo, its dedicated AI, data center, and cloud event.
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
1-3 December (Tuesday-Thursday): Black Hat MEA brings cybersecurity leaders, researchers, and technology companies to Malham, Riyadh.
6-12 December (Sunday-Saturday): NeurIPS holds its main AI research conference in Sydney, offering a look at advances that could shape the next generation of commercial AI.
7-11 December (Monday-Friday): GITEX Global opens with a leadership summit at Dubai World Trade Centre on 7 December, followed by its exhibition at Expo City Dubai on 8-11 December.
AI + Innovation (MENA) 
MENA+ Edition 
Egypt Edition 
UAE Edition 
Saudi Edition 
Infrastructure 
Share Issue
