US chip exports deepen AI ties with Saudi Arabia and the UAE

• The US cleared exports of up to 35,000 Nvidia Blackwell AI chips to G42 and Humain.
• The move deepens US–Gulf AI ties despite some lawmakers’ concerns.

The US Commerce Department has approved the export of advanced AI chips to two firms in Saudi Arabia and the United Arab Emirates. The authorisation covers the equivalent of up to 35,000 Nvidia Blackwell chips, which are central to large-scale AI training.

The two buyers — G42 in Abu Dhabi and Humain in Saudi Arabia — are working on major data centre projects in their countries. The timing of the approval matched the return of Saudi Crown Prince Mohammed bin Salman to the United States for the first time since 2018. It also signalled US support for the two governments’ AI ambitions.

In its statement, the Commerce Department said both companies can buy up to 35,000 Blackwell chips, though the total value can vary. A full batch would be worth about $1 billion. The agency said the approvals come with “rigorous security and reporting requirements.”

Earlier in the day, Humain said it plans to buy 600,000 Nvidia AI chips. Humain is also working with Elon Musk’s xAI to develop large-scale data centres in Saudi Arabia, including a planned 500-megawatt site.

The UAE praised the US decision. Ambassador Yousef Al Otaiba said the approval reflects ongoing talks between the two governments and shows their trust in each other on technology and security.

G42 aims to build one of the largest data centre hubs in the world using US technology. Nvidia, OpenAI, Cisco, Oracle and SoftBank are involved in its first phase, called Stargate UAE, which is set to go live in 2026.

New joint venture in Saudi Arabia

AMD, Cisco and Human also outlined plans to form a joint venture to support Saudi Arabia’s goal of becoming a major provider of AI services. The venture is expected to launch in 2026 and will combine Humain’s data centres with AMD and Cisco hardware. The companies plan to build up to 1 gigawatt of AI infrastructure by 2030.

The first phase will cover a 100-megawatt buildout that uses Humain’s data centre capacity, AMD Instinct MI450 GPUs and Cisco infrastructure.

Earlier this year, AMD and Cisco said they would work with Humain on what they described as an open and scalable AI buildout. The joint venture moves that plan forward.

AMD CEO Lisa Su said the three companies are bringing together compute and networking technology to help expand Saudi Arabia’s AI ecosystem. Cisco CEO Chuck Robbins said the project could play a key role in the country’s push to build an AI-driven economy. Humain CEO Tareq Amin said the partners aim to bring up to 1 gigawatt of high-performance infrastructure online in the Kingdom.

A new Cisco survey shows the gap the companies hope to fill. While most Saudi firms expect to use AI agents, fewer than one-third have strong GPU capacity. The report points to a need for more advanced data centre resources, which the joint venture plans to supply.

A race for better AI models

Major AI labs continue to release new models, even as investors look for signs that the rapid progress is translating into steady revenue.

Google introduced Gemini 3 this week, which outside evaluators say is among the strongest models available. The release comes 11 months after Gemini 2 and reinforces Google’s position in frontier research. At the same time, Elon Musk’s xAI pushed out Grok 4.1.

Still, new model launches often make headlines only when something goes wrong. Users are more focused on product experience and clear gains. Alphabet’s stock has risen this year as customers adopt its AI cloud services.

The competition for better models has also opened the door to more money. Microsoft and Nvidia have backed Anthropic with up to $15 billion, and Anthropic has committed to spend $30 billion on Azure.

That surge in funding aligns with the Saudi Crown Prince’s trip to Washington. His visit included a wide range of new US–Saudi investment deals, much of it touching on technology and AI.

MBS returns to Washington

Saudi Crown Prince Mohammed bin Salman’s visit to Washington was both political and commercial. At the US–Saudi Investment Forum, top US tech leaders — including Elon Musk, Nvidia’s Jensen Huang and executives from Google, IBM, Salesforce, Palantir, Cisco and Adobe — gathered with Saudi officials. The discussions highlighted how closely the two countries now work on AI and advanced tech.

For Silicon Valley, the Middle East has become a major source of capital. Many firms stepped back after the 2018 killing of journalist Jamal Khashoggi, but the region’s wealth has since drawn them back, especially after the tech market downturn in 2022. Investors began telling themselves they could work with less controversial funds, such as those in the UAE.

Ethical tension around Middle East funding

Some AI leaders continue to struggle with the question of who should back their work. CNBC reported that Anthropic CEO Dario Amodei resisted direct Saudi funding over ethical and security concerns. Yet Qatar’s sovereign fund later joined a major financing round. In a note obtained by WiredAmodei wrote, “This is a real downside… but ‘no bad person should ever benefit from our success’ is a difficult principle to run a business on.”

Sovereign funds in the UAE and Saudi Arabia are now major players in global AI investment. They want to diversify beyond oil and are using aggressive funding to buy access, partnerships and influence.

OpenAI’s proposed $500 billion Stargate data centre project has backing from MGX, an Emirati state-owned fund. OpenAI later announced plans for a data centre in Abu Dhabi to help governments build “sovereign AI capability” with US involvement. Nvidia, meanwhile, is lobbying to sell more high-end AI chips in the region.

Some US lawmakers are uneasy about this trend. They warn that advanced chips and models could spread to countries where US influence is mixed. The debate reflects the broader reality of 2025: AI sits at the centre of a global power struggle, and money from Riyadh and Abu Dhabi is reshaping the financial side of US innovation.

The environmental cost of the AI boom

As tech companies build more data centres to power AI systems, they are also trying to offset the environmental impact. Data from CDR.fyi shows rising demand for “durable” carbon removal credits — those linked to carbon stored for decades or longer.

Microsoft has bought more than 30 million such credits, far more than any other company. Google and Amazon are also among the biggest buyers. These purchases are helping drive up prices for durable credits tied to things like biochar, direct air capture and land restoration. They now cost nearly four times as much as traditional forest-based offsets.

But buying credits does not solve local issues tied to data centres, such as water use and strain on power grids.

More cloud discussions in Saudi Arabia as Saudi Arabia’s Public Investment Fund, the tech firm SITE and Microsoft signed an agreement to explore Microsoft’s sovereign-cloud services in the Kingdom, according to a post from the fund.

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.