The rules just changed for who gets the hardware. The United States finally cleared license-free exports of advanced AI chips and servers to the UAE.
It is a massive green light for infrastructure. Under the upgraded export control tier, the UAE government, tech giant G42, and approved US tech companies can now import these high-performance tools without waiting on bureaucracy.
“This strengthens the US-UAE AI acceleration partnership and accelerates development.”
This isn’t just a policy shift. It is physical capacity landing in Dubai. The US sees the region as a priority now. The barrier is down.
How the US chip deal impacts UAE AI infrastructure
The move directly targets the bottleneck every AI developer hates. Supply.
By removing license requirements, Washington allows Riyadh’s neighbor to stock up on the GPUs that power large language models. G42, the UAE’s state-backed AI powerhouse, leads the charge. But approved US firms get a ticket too.
It answers a question everyone was asking about US policy: Will we treat partners like rivals?
In this case, the answer is no. The tier upgrade means faster deployment. More servers. Less red tape. The result is a tangible speed-up for regional cloud infrastructure. You see more data centers. You see heavier computation. The groundwork for sovereign models is getting harder, faster.
Why sovereign Arabic AI is finally getting serious
Language barriers are falling, but only if you have the compute to back them up.
Canada’s Cohere just did two things that matter for Arabic speakers. First, they launched Transcribe Arabic. They call it the most accurate open-source speech recognition model available for the language. It doesn’t just read Standard Arabic. It handles the chaos of dialects. It catches when speakers mix Arabic and English, which happens constantly in the Gulf and North Africa.
Second, they are going to Saudi Arabia.
Cohere partnered with HUMAIN. Together they will build sovereign foundation models. Specifically for Arabic.
They aren’t promising air. They are backing this with at least 50 megawatts of dedicated compute power. It marks Cohere’s first expansion of physical compute outside North America.
This answers “where” for a lot of developers. You don’t need to route sensitive voice data to San Francisco to get good results. You can process it locally. This matters for government contracts. It matters for banks. Sovereign data stays sovereign.
Saudi Arabia and Egypt lead in governance and risk management
Building models is fun. Managing them is boring and necessary.
Saudi Arabia dropped a new national framework. The Saudi Data and Artificial Intelligence Authority (SDAIA ) issued rules to manage AI risk. It covers seven specific risk categories. From identifying the problem to treating the outcome, they have a consistent path now for both government entities and private industry.
Meanwhile, Egypt is getting broader support.
The World Bank pledged help for Cairo’s digital ambitions. They aren’t just giving advice. They are supporting the implementation of the second National AI Strategy. This includes broadband expansion, which is the unsexy prerequisite for everything else. If the internet is slow, AI is a hobby. If the internet is fast, AI is infrastructure. The discussions covered data governance too. They know the game isn’t just about speed anymore.
Jordan also chipped in with money.
Jordan’s Innovative Start-ups and SME Fund (ISSF) threw $5 million at Saudi Technology Ventures’ AI fund. It creates a pipeline. Capital and expertise flow from the bigger economy into Jordanian start-ups.
Real-world wins in farming, sports, and banking
Most AI news sounds theoretical. This week had sweat equity.
Haswah Mango Farm in Saudi Arabia used AI to look at the ground. They monitored soil moisture. They watched water quality. They tracked crop health.
The result? 50% less irrigation water.
And production went up 40% for the 2026 season? No, the report says “at least 40%”. The technology isn’t magic. It’s just eyes that never blink and brains that process data faster than a farmer can walk the fields.
In sports, Amplifai launched NUR. It uses thermal imaging to watch athletes recover. It tells doctors and teams if a player is ready to return to the game. It reduces guesswork in return-to-play decisions. Early deployments are already in universities and sports clinics.
Banking saw similar concrete steps. Emirates NBD in the UAE became the first bank in the region to use the Partior network for blockchain-based US dollar payments. It cuts down on time and adds transparency for corporate clients.
In Egypt, the Central Bank started its fifth cohort of a Generative AI in finance program. They are training people. Large language models and financial modeling. It builds skills inside the bank.
Funding rounds and startup focus areas
Money is moving to the edges of the map.
Think, an infrastructure play out of Riyadh, raised $8 million. Pre-seed stage. Investors include RAED and Wa’ed. They claim their platform keeps GPUs at over 90% utilization. Industry averages sit at 30-50%. That difference eats into profits. Fix the waste. Lower the cost. Simple math.
In construction, Kuadra is building Egypt’s first unified platform for that sector. They automate tenders and procurement. Days of document processing become hours. They plan to hit oil and gas next.
Orange Jordan kicked off its Summer Challenge again. Students and start-ups pitch solutions. Mentors judge. The focus is always the same. Business growth via automation.
Culturally, SDAIA signed a deal with Saudi’s Ministry of Culture. They are looking at all 16 recognized sectors. Policy first. Adoption second.
Awards shifted focus too.
The UAE National AI Award added Agentic AI. This isn’t a side note. It reflects a strategic pivot toward agents that act, not just chat. Sharjah followed suit. Their communications awards added AI categories.
Dubai Media launched “Media X.” An AI transformation program for their whole content ecosystem.
None of this is wrapped up in a neat bow.
The chips are here. The frameworks are drafted. The models speak dialect. But the heavy lifting? The integration into the messy daily grind of government, farms, and banks?
That is where the next quarter is going to be spent.
