Six Arabic AI Openings in Saudi Arabia’s Year of AI

Key findings

In the Kingdom’s Year of AI, the obvious Arabic AI ideas are already being built. The less obvious ones are still open.

1. 27 products and initiatives already compete in six “obvious” Arabic AI lanes in Saudi Arabia, from dialect customer-service agents to legal and clinical tools.
2. 6 lanes are open: in five, our scan found no dedicated Saudi AI product; in the sixth, it found one early-stage regional entrant. They include Arabic government-tender drafting, official correspondence and bilingual technical documentation.
3. $219 million across 72 deals: Saudi startup funding in H1 2026, down 74% year on year, with fintech taking 67%. AI application builders are competing for a smaller pool.
4. 55% of the capital deployed by MENA venture funds launched in 2026 went to Saudi startups, a signal that new money is arriving.
5. 300 active data and AI startups by 2030 is the target set in the 2020 National Strategy for Data and AI. Open lanes are where the next ones can come from.

Six lanes still open

In five of these lanes, our scan did not identify a dedicated Saudi AI product as of 19 October 2026; in the sixth, it found one early-stage regional entrant. All six share three traits: the work is in Arabic, it is repeated every week, and getting it wrong has a visible cost. That makes them a natural fit for a model built for Arabic and deployable inside the Kingdom.

1. Government tender responses in Arabic. Tender documents submitted to Saudi government entities must be in Arabic, according to procurement guidance (OpenOpps). Companies still recruit bilingual coordinators to monitor Etimad and translate bids by hand (job listing). Opportunity: bid drafting, compliance checklists and bilingual consistency. Caution: final submissions need human sign-off.

2. Official correspondence and protocol. Letters to ministries and officials follow strict forms of address and titles. The tools we found are general writing assistants, such as Kaatib’s formal-email feature (listing). Opportunity: template-and-rules drafting for HR and government-relations teams. Caution: titles must be exact, so free-form generation alone is not enough.

3. Bilingual technical documentation for a Saudized workforce. Saudi Aramco’s contractor guide notes that Saudization percentages in most of its contracts rise each year, and bilingual documentation supports that workforce development (guide). Industrial employers advertise for translators who maintain technical term databases (a job listing on Sabbar, since removed). Opportunity: versioned Arabic–English procedures with an approved terminology base. Caution: safety-critical content needs engineer approval.

4. Sharia governance documentation. One early-stage regional entrant, Daeson Technologies’ Shariah AI Compliance Co-Pilot, was introduced in March 2026 (report). Certification brings recurring work: Syarah’s Sharia certification includes annual audits (Sharikat Mubasher). Opportunity: clause extraction and documentation packs for Sharia boards. Caution: rulings remain with the board.

5. Local content documentation for suppliers. Market-entry guides note that qualifying government tenders require a Local Content Certificate from LCGPA (guide). Buyer-side tools are emerging, such as Penny’s partnership with SAP (CPOstrategy). Opportunity: supplier-side evidence packs in Arabic. Caution: scoring rules are set by LCGPA.

6. Specialised agents for HUMAIN ONE. HUMAIN said in May 2026 that HUMAIN ONE will be available on AWS Marketplace (HPCwire), and in March 2026 announced a marketplace for third-party agents. Opportunity: niche Arabic agents built for that channel. Caution: commercial terms for third-party agents should be confirmed with HUMAIN.

The setting: a national push, a tighter funding market

In March 2026, the Saudi Cabinet designated 2026 the Year of Artificial Intelligence (Argaam). The Kingdom ranked 14th in Tortoise Media’s Global AI Index in its 2024 edition (Argaam).

The National Strategy for Data and AI, launched in October 2020, set three 2030 targets that frame this report: 300 active data and AI startups, 20,000 data and AI specialists, and more than $20 billion in investment (Arab News).

The funding picture is mixed:

MAGNiTT attributes much of the Saudi decline to the absence of large late-stage rounds rather than an even slowdown across the market. For AI application builders, the practical reading is that capital is selective, and a clearly defined, under-served lane matters more than ever.

Funding fell by nearly three quarters while deals fell by two fifths, so the average round got smaller.

The new money is tilting toward the Kingdom: of the $60M deployed by funds launched in 2026, the largest share went to Saudi startups.

What founders can build on today

ALLaM, the Arabic-first model family that began at SDAIA’s National Center for AI and is now carried forward by HUMAIN, is reachable through four public channels. They suit different builders.

Three details stand out for builders:

• Adoption is already measurable. The 7B model’s Hugging Face page lists about 59,000 downloads in the last month, 30 fine-tuned derivatives and 28 adapters (snapshot, 19 October 2026).
• Data rules favour local deployment. Saudi Arabia’s Personal Data Protection Law regulates cross-border transfers of personal data, so in-Kingdom hosting may simplify compliance for many enterprise workloads. Companies should take their own legal advice.
• A new distribution channel is opening. HUMAIN announced in March 2026 a marketplace on HUMAIN ONE where third-party developers can publish and sell AI agents (Arab News).

Six lanes where builders are already active

Our scan identified 27 products and initiatives active in the Saudi market across the six lanes most often proposed for Arabic AI. Their presence is a sign of demand, not saturation: a new entrant here needs a clear point of difference.

Two data points show how far these lanes have matured. Nuxera reports clinical documentation across 28 Arabic dialects, and Clauze.AI launched in September 2026 with a strategic investment from White & Case.

The list is illustrative, not exhaustive. Inclusion reflects public information only and is not an assessment of any product’s quality or market share.

Where national platforms moved first

In two lanes that often appear on startup idea lists, national platforms have already deployed AI at scale. For founders, the better role here is supplier or integrator, not competitor.

• Building permits. The Ministry of Municipalities and Housing issued more than 34,000 building permits in H1 2026, with its Balady platform handling applications digitally (Amlak). An automated design-review engine integrated with Balady is described in an industry award submission (Stevie Awards entry).
• Motor insurance claims. Najm, the national motor-accident platform, and Solera announced a partnership, reported in August 2026, to deploy AI estimating and claims tools across Najm’s platform (Middle East Insurance Review).

What it means

Ready to quote

Attributed to the Made in 10 Research Team:

Methodology

• Period: desk research conducted 19 October 2026.
• Scope: Arabic-language AI products and initiatives active in the Saudi market, grouped into 14 lanes commonly proposed for Arabic AI: 6 active, 2 led by national platforms, and 6 open.
• Method: structured web searches per lane in English, followed by review of company, government and media pages. A lane counts as “open” when no dedicated Saudi product was identified for that specific use.
• Figures: every number is taken from the linked source. Funding data is MAGNiTT’s; download counts are a single-day snapshot.
• Limitations: desk research can miss companies, especially those without English-language coverage, and markets change quickly. Counts are a lower bound. An “open” lane means none was found, not that none exists.

Disclosures and corrections

• Made in 10 is a venture builder that builds products. It may develop products in lanes described in this report.
• This report is independent research. It is not affiliated with, sponsored by or endorsed by SDAIA, HUMAIN or any company named. All names and trademarks, including ALLaM and HUMAIN ONE, belong to their owners. Neither HUMAIN nor SDAIA has reviewed this report.
• Company mentions are based on public information and are not endorsements, ratings or assessments of quality.
• This report is for information only and is not investment, legal or financial advice.
• Companies that believe they are missing or misdescribed can write to research@madein10.com. Verified corrections will be published in an updated version.

Suggested citation

Made in 10 Research Team (2026). Six Arabic AI Openings in Saudi Arabia’s Year of AI. Report MI10-R-26-01. madein10.com/research

Charts and tables may be reused under CC BY 4.0 with attribution to the Made in 10 Research Team.