The AI race is typically seen as a US–China showdown. Yet the UAE, a small country with outsized AI ambitions, is carving out a sphere of influence in open-source strategy. Take a look at this leaderboard on Artificial Analysis for the top open-source LLMs (Large Language Models). There are the UAE’s homegrown AI models. Falcon was the first sign of the UAE’s rapid rise in the world of AI.

At a time when leaders like OpenAI and Google DeepMind lock their top proprietary models behind paywalls, the UAE is betting on openness. Its AI models are free to use and modify, and are designed to run on less computing power, lowering technical barriers and expanding who can actually use them. 

This move matters. By making advanced AI accessible, the UAE is expanding its innovation base—enabling researchers, startups, and public institutions to build without prohibitive costs—while simultaneously positioning AI as a strategic national asset and a tool of soft power.

Research firm Forrester has stated that open source can help firms “accelerate AI initiatives, reduce costs, and increase architectural openness,” ultimately leading to a more dynamic, inclusive tech ecosystem.

In the UAE, two key organizations, Mohammed Bin Zayed University of Artificial Intelligence (MBZUAI) and Technology Innovation Institute (TII), are leading open-source AI projects and often outperforming Meta’s and Google’s top offerings. TII created the Falcon Series to expand generative AI across more areas, including Falcon Arabic, which captures the full linguistic diversity of the Arab world.

Falcon models cover LLMs, Edge AI, and Arabic-optimized versions. Falcon supports 18 languages and is widely adopted worldwide, including Europe, Latin America, Africa, and the Middle East, for applications in healthcare, energy, and more. TII is even exploring the use of Falcon models to control robotic agents, signaling the convergence of AI and robotics as the next frontier of application. 

When the Falcon 2 series was launched in 2024, Advanced Technology Research Council (ARTC) Secretary General Faisal Al Bannai told Reuters: “We’re very proud that we can still punch way above our weight, really compete with the best players globally.”

In 2024, ATRC pledged $300 million to the Falcon Foundation to support ongoing open-source work on the LLM series.

Then in 2025, MBZUAI and G42, the UAE’s leading AI company, launched K2 Think, a small but powerful and novel reasoning system. At 32 billion parameters, K2 Think performs faster in math reasoning than models that are an order of magnitude larger, such as DeepSeek R1. This system adds to the UAE’s growing group of open-source models, such as Jais, which is considered the world’s most advanced Arabic large language model. 

In January, G42, Cerebras, and MBZUAI’s Institute of Foundation Models released K2 Think V2, a 70-billion-parameter reasoning system widely seen as a milestone in the UAE’s push for technological sovereignty. Fully open by design, the model allows industries—from healthcare to aerospace and finance—to fine-tune it on secure, in-house data without exposing sensitive information to third-party clouds.

Transparency is central to K2 Think V2. MBZUAI has released the full training stack: pre-training data and curation details, intermediate checkpoints that capture the model at different stages of learning, and the exact instructions used to build it. This level of openness enables inspection, auditing, and reproducibility in ways proprietary models such as GPT-5 do not.

At the same time, the country is forging deep partnerships with big tech firms—not only to integrate into global value chains but also to maximize technology transfer. G42 has partnered with OpenAI to deploy advanced AI solutions across the UAE and the region, and signed a $100 million deal with Cerebras to build what it calls the world’s largest AI training supercomputer.

SHOWING RESULTS

The UAE has been investing billions of dollars to become a global AI leader, and building a vibrant open-source ecosystem is already showing tangible results—from real-world use to attracting top AI researchers.

“K2 Think shows that open systems can move rapidly from research to real-world use, delivering top-tier reasoning performance today,” says Richard Morton, Acting Executive Director of the Institute of Foundation Models at MBZUAI.“Building on this, our first K2 Think Hackathon demonstrated how this momentum is translating into global, real-world adoption, attracting more than 3,000 applicants and around 900 applications from over 50 countries, including participants from leading universities such as Harvard, MIT, and CMU, as well as institutions like DeepMind, McKinsey, Alibaba, and the Mayo Clinic.”

The resulting use cases, Morton adds, spanned more than 10 sectors, from healthcare and education to finance, and engaged over 500,000 target users worldwide. “As adoption grows across academia, industry, and government, the multiplier effects of open-source AI will continue to scale over the coming years.”

Driving real-world impact, Falcon models, in partnership with the Bill & Melinda Gates Foundation, support AgriLLM—helping farmers make smarter decisions amid climate stress.

While many AI models chase narrow consumer use cases, TII has focused on building foundational models built for scale: adaptable to industry, research, and public-good applications, without sacrificing accessibility. 

“We are already seeing tangible results, particularly in model performance, real-world use cases, and community access,” says Dr. Hakim Hacid, Chief Researcher of the Artificial Intelligence and Digital Science Research Center at TII. “Open access models are enabling practical applications today.”

He adds that the full multiplier effect—widespread adoption, ecosystem growth, and downstream innovation– will unfold over several years as developers and institutions build on these open foundations.

PATH TO INNOVATION AND DIGITAL AUTONOMY

For the UAE, an open-source strategy is both an engine of innovation and a path to digital autonomy.

“Open foundations make it easier to experiment, localize, and deploy AI while retaining control over governance and use,” says Hacid. “Rather than chasing scale alone, the focus is on efficiency, architecture, and adaptability—offering a complementary path to competitiveness built on openness, collaboration, and real-world usability.”

Morton says that open-source AI aligns closely with the UAE’s vision of inclusive technology and broad societal benefit. A truly open approach, he adds, demonstrates that “frontier-class capabilities can be achieved through transparency and community-driven innovation.”

With the government prepared to commit greater funding and compute, successor models are expected to be significantly more powerful. The question now is whether the UAE can translate this momentum into open-source AI leadership on par with the US and China.

“The UAE is becoming a credible and consistent contributor to global open-source AI, and the trajectory is encouraging,” says Hacid, adding that by focusing on areas such as efficient model design, Arabic-language capabilities, and open scientific practices, the UAE is building distinctive strengths.

“Leadership in open source is about sustained contributions, adoption, and impact.”

Launched in 2025, IFM is built to operate as a distributed organization across Abu Dhabi, Paris, and Silicon Valley. By open-sourcing not only its models but entire development pipelines, MBZUAI positions IFM as a leader in open AI development. Through the LLM360 initiative, researchers gain full access to training code, datasets, and model checkpoints. 

“We believe models like K2 Think are industry-leading and highly competitive, placing them in the top tier of AI systems globally,” says Morton.

Open source gives everyone permission to innovate, without installing any single company as a gatekeeper. Experts believe that this same spirit of open innovation will democratize advanced AI capabilities in the UAE.

“Open models accelerate scientific progress, improve transparency, and enable local adaptation, ensuring AI development benefits have the widest impact instead of benefiting a small group of proprietary providers,” says Morton.

As AI becomes central to how we learn, work, and decide, the UAE’s open-source push is about more than technology. It’s about building distributed communities of innovation—and ensuring the benefits of AI are shared more broadly across sectors and society.