E-Paper
Sign In
Sun, Nov 09, 2025 | Jumada al-Awwal 18, 1447 | Fajr 05:12 | DXB 
28.4°C
iPhone | iPad
Android
AI in Arabic? How Gulf could soon lead Artificial Intelligence race
What began as a push to digitise economies has become a contest to build and even export intelligence that mirrors the region
The Gulf’s AI ambitions are getting louder and more local. At the recent GITEX Global 2025, MENA’s biggest tech showcase, the well-deserved spotlight was on homegrown AI. Startups like Arabic.AI demoed large language models trained to think, read and respond in Arabic, proof that the region’s race to build its own intelligence is past the talking stage.
Across the UAE and Saudi Arabia, the focus has shifted from adoption to creation. New systems are being built to process regional data and nuance, not just English text. What began as a push to digitise economies has become a contest to build and even export intelligence that mirrors the region. “We’re already seeing Arabic large language models integrated across finance, government services and customer support,” says Bashar Alhafni, Assistant Professor in Natural Language Processing at the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI). Alhafni, who leads Aram, the university’s Arabic AI Modeling Lab, studies how Arabic language technologies can be built around human context so that systems are linguistically precise and socially grounded. He points to major Arabic LLMs such as Jais from the UAE, ALLAM from Saudi Arabia, and Fanar in Qatar. “The combination of strategic policy and robust infrastructure has made the GCC the primary engine for Arabic-first AI development,” he says. That foundation rests on three pillars: clear government direction, serious infrastructure and technical talent, with efforts to advance technologies beyond labs and integrate them into broader economic strategies.
The Gulf advantage
Recommended For You
Google’s 2024 MENA Economic Impact Report found that AI tools generally added Dh21.8 billion to the UAE economy while a Crowell & Moring analysis projects regional tech spending could reach $169 billion by 2026. In Dubai, the State of AI Report found that nearly three-quarters of organisations are already investing in Arabic-language or localised systems — a sign that building “in Arabic” is becoming a business strategy.
Simone Vannuccini, Professor of Economics of AI and Innovation at University Cote d’Azur studies how policy affects deployment of intelligence across industries. “I think the Gulf countries are positioning themselves as a ‘middle-way’, alternative platform for AI compared to Chinese and US trajectories,” Vannuccini says. “This approach may seek to present itself as a ‘global hub’ for AI, or at least a regional one, while the rest of the world seems to close itself within borders.”
Alhafni refers to the growing global market for Arabic AI, within and beyond the Arab world. Governments and universities are looking to license or integrate Arabic LLMs for education, translation and cross-cultural communication. While the interest isn’t new, today’s scale and capability have turned it from an academic pursuit into a real opportunity.
Sonny Arcot, founder and CEO of Arcot Group, is one of several entrepreneurs betting on the Gulf’s AI momentum. The company operates across the UAE, the US, and India, offering several AI products that automate workflows in various sectors.
Arcot’s focus is on solving language and compliance gaps. The company’s AI offerings are designed to help organisations process data, personalise user experiences, and improve compliance, areas where most off-the-shelf systems built for English markets still fall short. Arcot chose Dubai as its base because of the UAE’s forward-thinking approach to technology and regulation. “Just like the U.S. has Silicon Valley, I believe that shift is happening here,” he says. While his goal is to make AI adaptable in the region across languages and industries, he believes a technology’s ability to process Arabic-language data securely and contextually will determine a company’s success.
From research to revenue
For governments, that next phase — scaling AI beyond national projects — is where firms like PwC Middle East come in. Moussa Beidas, the firm’s Partner and Ideation Lead, advises the UAE and Saudi Arabia on how to turn national strategies into global workable ecosystems. Beidas believes both countries are well-positioned to lead Arabic AI because they combine capital, mandates, and data and tie in their national AI strategies into economic transformation plans. Pointing to Saudi Arabia’s goal of becoming a regional AI hub, he says, “We work to bridge linguistic and cultural gaps, ensuring that LLMs and other AI systems can generate nuanced Arabic content.”
When it comes to adoption, government agencies remain the main testing ground, powering citizen services, digital ID systems and smart-city projects, with banks, media and entertainment space catching up. Beidas says sectors like healthcare, education, and SMEs still face hurdles though at MBZUAI, they’re addressing this through ARWI [Arabic Write & Improve], an AI-powered Arabic writing assistant designed to provide pedagogically aligned feedback to students. That domestic focus mirrors what PwC’s 2025 AI Jobs Barometer found: AI-related roles have surged but will require building and upskilling local talent to design, adapt, and govern systems from the ground up. “The real test,” says Beidas, “will be whether that growth can extend beyond national markets.”
The UAE’s AI growth is accelerating, but most of the activity is still happening inside national borders. Turning Arabic-trained systems into exports — models, APIs, or enterprise tools that work across the region — will depend on how well neighbouring countries align on data, licensing, and infrastructure. That’s where experts see the next phase of competition: exporting Arabic AI that works across borders.
Exporting intelligence
Alhafni believes that regional adoption of GCC-trained Arabic models will require policy alignment across countries, infrastructure for model deployment beyond the Gulf, and dialectal coverage reflecting linguistic diversity. Vannuccini agrees. He says there is a space for ‘regional models’ becoming a go-to solution for certain areas of the world like the MENA region. “Let’s imagine large public hospitals or universities want to adopt AI-based systems, they may turn towards solutions that are ‘regional’ because they are more proximate in the geographical, cultural, and language space,” he says.
That kind of regional openness within AI, says Beidas, could turn the Gulf’s AI edge into a broader industry. Alhafni says, “Jordan and Lebanon have great engineering talent, lower costs, and multilingual markets, making them ideal for refining Gulf-built models. It creates a healthy regional value chain — the Gulf leads on large-scale development, while the Levant helps localise and commercialise those systems.”
The Gulf still holds the biggest short-term opportunity for Arabic AI, with Egypt close behind. Beidas feels that Levant countries and North Africa could support, helping fine-tune Gulf-built systems for different dialects and local markets. But closing the region’s AI gap will require steady investment in talent, infrastructure, and cross-border collaboration.
That cooperation is already taking shape across the private sector. Andrew D’Souza, CEO of Boardy, described his company as an “AI super-connector” that turns short conversations into curated, double-opt-in introductions for founders, operators, and investors. Boardy works with Gulf-based partners and events to help link ecosystems through applied AI. “From our vantage point, the UAE’s strength lies in its world‑class infrastructure and ambition,” says D’Souza, adding that the UAE has the infrastructure, ambition, and funding figured out but is missing the “last mile” between innovation and real-world products. The real potential, he says, is in scaling that kind of collaboration across MENA through curated partnerships.
A huge hurdle is usable Arabic data. Most existing datasets still lean towards a handful of dialects or regions. To grow responsibly, experts say, AI needs to become more inclusive and trained on data reflecting the Arab world’s linguistic range.
Even with those gaps, Beidas believes interest in Gulf-built models is rising fast, especially among public-sector and financial institutions in Arabic-speaking countries that want to localise services. What’s still missing, are standardised licensing systems, stronger data-sharing frameworks, and clearer commercial returns to prove the business case.
Vannuccini believes the main bottleneck is demand. “If AI does not show real potential beyond the hype in a relatively short time, final demand might collapse or jump to the ‘next big thing’,” he says.
Forecasts ahead
Looking ahead, the region’s trajectory is clear. “GCC-trained Arabic models are maturing rapidly, and as they continue to improve in dialectal coverage, efficiency, and user alignment, they’ll start powering applications across education, government, and commerce,” says Beidas. The next five years, will be about turning prototypes into real-world impact.
Others, like Marcello Mari, believe that exportability is already happening, albeit not in the traditional sense. “AI is already exported as cloud services, on-prem deployments, and open-licensed model weights,” says Mari, former CEO of SingularityDAO. “Ownership depends on licensing, not geography. Models under MIT or Apache licenses grant broad rights, while others restrict redistribution or commercial use.”
For Mari, the conversation is about who controls AI. His work on decentralised systems which are designed to distribute data access and decision-making, parallels the Gulf’s push for digital sovereignty. Initiatives like the UAE Strategy for AI and Saudi Arabia’s National Strategy for Data and AI reflect that goal: keeping data, talent, and economic value inside the region.
He cites the UAE’s Jais model and Saudi Arabia’s ALLAM, as examples of that ambition in motion. Both projects were designed to create Arabic-trained systems that reflect local context reducing dependence on foreign AI.
“The main constraint for developers in the Arabic-AI space remains the lack of high-quality Arabic data,” Mari says. The UAE and Saudi Arabia are now investing heavily to close that gap through new datasets, evaluation tools, and training programs. The next stage of growth will depend less on new technology and more on what surrounds it: open licensing, affordable data pipelines, and transparent evaluation methods.
Beidas sees a similar trajectory. “We’re already watching a regional ecosystem take shape,” he says. “The Gulf builds large-scale models, while countries like Jordan and Lebanon help fine-tune and commercialise them. It’s a healthy value chain.” But real exportability will require shared datasets, performance benchmarks, and coordinated regulation — “the connective tissue” that turns policy into markets.
Realistically, it may take a three-to-five-year timeline for the Gulf to start exporting Arabic AI through licensed models, APIs, and enterprise tools, as regional data ecosystems and partnerships mature. Vannuccini believes in the long term, the real value capture being for those that have stacks in datacenters and physical infrastructure. He says: “While everyone looks at OpenAI, it is Nvidia that has passed from being a gaming hardware company to a top worldwide giant. The Gulf prospected investments in France or UK go in this direction.” Still, if the region succeeds in exporting Arabic-trained models that reflect its own culture, language, and governance, it won’t just be creating tools, it will be exporting identity in a way. A homegrown technology signals the Arab world’s step back from being consumed by imported technologies built elsewhere. And that’s the quiet revolution inside this AI race: the move from participation to authorship.
Don’t just scroll. Know it all. Sign up for your daily UAE news fix.
Get the top stories of the day right in your inbox.
