SDAIA AI Training Programmes: How Saudi Arabia Is Teaching a Million People AI
1.56 million beneficiaries, 1.2 million SAMAI completions, 14,495 specialists: what SDAIA’s training numbers really measure, which programme suits whom, how to apply, and what a trained workforce changes for companies deploying AI.
Saudi Arabia’s AI training effort is usually reported as one big number, and the number changes every time. One million. 1.1 million. 1.2 million. 1.56 million. All of these are real, and they measure different things. If you are deciding whether to enrol, or planning to hire the people who did, the difference matters more than the headline.
This guide sets out which SDAIA training figures hold up as of September 2026 and which programmes exist that you can verify. It covers who each one is for, how to apply, what it costs and what you actually learn. Then it turns to the organisational question: what a trained workforce changes for a company deploying AI, and what no course can do for it.
The numbers, checked one by one
In June 2026 the Saudi Data and Artificial Intelligence Authority (SDAIA) said more than 1,563,983 citizens had benefited from its training programmes, 14,495 of them data and AI specialists and experts. It also said more than 278 government leaders had been trained. That figure is genuine and reported by the Saudi Press Agency. But it covers all of SDAIA’s training, and the word is “beneficiaries”, not graduates.
The figure most people actually mean is the one for SAMAI, the “One Million Saudis in AI” initiative. It passed its one-million target in November 2025. At a press conference during the UNESCO Global Forum on the Ethics of AI in Riyadh on 14 September 2026, SDAIA President Dr Abdullah Alghamdi put SAMAI completions at more than 1.2 million.
All bars are people, on one scale, using figures published by SDAIA, the Ministry of Education and the Saudi Press Agency. The two smallest are drawn at a minimum visible width; at true scale they would be close to invisible, which is rather the point.
Read the bars as a pyramid, not a sum. The wide base is AI literacy, which means using tools sensibly, understanding what a model is and handling data responsibly. The narrow top is the 14,495 people SDAIA counts as specialists. Both matter, but they answer different hiring questions, and a planner who confuses them will expect far more engineering capacity than the numbers support.
How the programmes fit together
SDAIA’s training is not one scheme but a family that has grown over five years. Our overview of the 2026 Year of AI treated the one-million figure as a single signal. The detail is in the sequence.
- 2021SDAIA Academy launchedFree programmes for ages 18 to 60, including specialist capacity building, awareness courses and cooperative training for university students.
- May 2023Elevate begins with Google CloudA free programme for women, with a first phase of 1,000 trainees from 28 countries and a five-year aim of about 25,000.
- Sep 2024SAMAI launched at GAINSDAIA with the Ministries of Education and of Human Resources and Social Development, aiming at one million citizens.
- Nov 2025One million passedThe three partners mark more than one million citizens trained and certified.
- Jan 2026Athka X, SAMAI 2 and a university curriculumAnnounced at the ICAN 2026 capacity-building conference, along with agreements with 14 universities.
- Jul–Aug 2026Summer bootcampsSDAIA Academy runs beginner and professional bootcamps, registered through Athka X.
- Sep 20261.2 million SAMAI completionsAnnounced in Riyadh, with a sector-focused SAMAI 2 described as the next phase.
The shape is deliberate. SAMAI built a broad base quickly. The Academy and Athka X handle depth. SAMAI 2 and the university curriculum aim the effort at people already inside institutions, namely civil servants and every undergraduate, rather than at the general public.
SAMAI: what it is and what you actually learn
SAMAI was launched in September 2024 at the GAIN Summit in Riyadh by the Minister of Education and the SDAIA President. It runs jointly with the Ministry of Human Resources and Social Development. Its stated aim was to teach a million citizens what AI is, how it is used and what its ethical dimensions are, and to train them on the tools needed to apply it at work and in daily life.
According to SDAIA’s public calls for registration, it is free, delivered in Arabic and remotely. It is open to all citizens, and it ends with an SDAIA-accredited certificate once the training hours are complete. Registration has run through a page on FutureX, the National eLearning Center’s platform. The people who took part were not mainly students: of the first million, 70% were employees and 30% students, and 52% were women.
On content, the most useful public description came in September 2026 from SDAIA’s head of capacity building. He said a significant part of SAMAI dealt with ethical and responsible use: which data is safe to put into a tool, which tools are safe to use, and how to protect personal and organisational data. SDAIA builds ethics into its courses rather than teaching it as a separate subject.
The next phase, SAMAI 2, was announced in January 2026 as a programme for government employees, developed with 11 ministries including Health, Finance, Justice, Education and Energy. In September 2026 SDAIA described it as sector-focused training in energy, health, industry, education and agriculture, aimed at producing practitioners in each field. The details of who can enrol, and how, were still emerging at our publish date.
SDAIA Academy and Athka X: the professional end
If SAMAI is the base of the pyramid, SDAIA Academy is the route upward. It was launched in 2021 for ages 18 to 60, is free to register, and has run four programme strands. These are national capacity building for data-science specialists, a data and AI awareness programme, a programme for organisations that use SDAIA’s services, and cooperative training for university students.
Since January 2026 the front door has been Athka X, which SDAIA calls the Kingdom’s first integrated national platform for data and AI capacity building. It lists programmes and bootcamps run with technology companies and universities, in person, remote, blended or self-paced. Each learner keeps a national skills record. It is aimed at students, graduates and working professionals.
The summer 2026 bootcamps show what the professional end actually teaches. Tracks announced in July covered data engineering for AI systems, computer vision for developers, AI systems and agent development, generative AI solutions and advanced agent engineering. Beginner tracks covered AI for productivity, prompt engineering, responsible AI use and generative coding. An August set of 18 bootcamps added data management, governance, quality and integration. Before that, in December 2025, the Academy opened a four-week professional programme on large language models that prepares trainees for an NVIDIA generative-AI certification exam.
That list is worth reading closely. Data governance and data quality appear as their own tracks next to agents and computer vision. SDAIA is teaching the unglamorous half of AI deliberately, and organisations should hire for it just as deliberately.
Universities, schools and the cooperative route
Two changes in 2025–26 move AI training out of opt-in courses and into the education system itself.
At ICAN 2026 in January, SDAIA launched a national cross-disciplinary curriculum in data and AI for all undergraduates, whatever their subject. It does not by itself make a course compulsory. It gives universities a ready-made basis for a mandatory data and AI course in every programme, built around using AI tools responsibly in study, work and decisions. SDAIA signed agreements with 14 public and private universities at the same event. In September, Dr Alghamdi said more than 40 academic AI programmes had been introduced since the start of 2026.
In schools, the National Curriculum Center announced an AI curriculum across public education starting in the 2025–26 academic year. It is being developed with the Ministry of Education, the Ministry of Communications and Information Technology and SDAIA.
For students who want more than a course, the cooperative training route is the one to know. In June 2026 SDAIA began training 215 students from 28 universities at its Riyadh headquarters, combining instruction with work on its own projects. Real projects inside a working data organisation are the closest thing to experience that a student can put on a CV.
Which programme suits which person
Most people need one starting point and one next step, not the whole catalogue. Here is how the verified programmes map onto four common situations.
| Student | Career switcher | Government employee | Developer | |
|---|---|---|---|---|
| SAMAI (general literacy) | Yes | Yes | Yes | No |
| SAMAI 2 (government and sector) | No | No | Yes | No |
| Academy beginner bootcamps | Yes | Yes | Yes | No |
| Academy professional tracks (agents, vision, LLMs) | Final year | After basics | Technical roles | Yes |
| Data governance and quality tracks | No | Yes | Yes | Yes |
| Cooperative training at SDAIA | Yes | No | No | No |
| University data and AI course | Yes | No | No | No |
This is our reading of each programme’s published audience, not an official eligibility table. Cohorts, dates and entry criteria change, so check Athka X or the programme page before planning around any row.
Read in practice, the table gives four paths. A student should take the university course if their institution offers it and SAMAI for the certificate, then aim for cooperative training. A career switcher should start with SAMAI, move to a beginner bootcamp, and then pick either a technical track or a governance track. The governance route is less crowded and often closer to what employers lack. A government employee should watch for SAMAI 2 through their ministry. A developer can skip the literacy layer and go straight to the professional tracks. The LLM programme is the one tied to an external certification.
How to apply, and what it costs
The verified picture on cost is simple. SAMAI is free, according to SDAIA’s registration calls. So is SDAIA Academy registration, according to the Academy’s published profile, as was Elevate. For individual Athka X bootcamps the announcements we read did not state a fee or eligibility rules. Treat any bootcamp as unconfirmed on both points until its page says otherwise.
- 1Start with SAMAIFree, in Arabic and online, with an SDAIA-accredited certificate. It gives everyone a shared vocabulary, including the people who will later manage engineers.
- 2Create an Athka X profileProgrammes, schedules and the national skills record live there. Registration opens per cohort, so check regularly rather than once.
- 3Pick one track and finish itEngineering, computer vision, agents or governance. One completed track is worth more on a CV than three you started.
- 4Get real-project exposureCooperative training for students, or an internal project at work for employees. Training without a live system to operate fades quickly.
One caution on certificates. Third parties can only check them if they can see what was assessed. When you list a course on a CV, say what you built or did in it, not just its name. Employers are learning to tell a literacy certificate from a technical one, and they will read past the logo.
What a trained workforce changes for companies deploying AI
For an organisation buying AI, the training figures matter less as national pride and more as an operating assumption. Three things change.
The handover gets easier to plan. The quiet failure in many AI projects comes after go-live. The integrator leaves, and nobody inside can update the knowledge base, read the logs or tell a model problem from a content problem. A pool of locally trained practitioners, including governance-track graduates, makes it realistic to staff that role from inside rather than renew a support contract forever.
Imported operators become less necessary. Not unnecessary, because 14,495 specialists across a whole economy is not many. But the case for flying in a team to run a routine assistant is weaker than it was. That changes total cost, and it changes who holds the system’s knowledge.
Staff use AI more safely, and more of them use it. Seventy per cent of SAMAI’s first million were employees, so a meaningful share of your current staff may already have had basic training in what data should not go into a public tool. That makes an approved internal tool easier to roll out, and shadow use of unapproved ones easier to discuss honestly.
What training cannot replace
A trained workforce is necessary and not sufficient. Several things that decide whether an AI deployment is safe and useful sit outside any course, because they are properties of your organisation and your system, not of your staff.
| Training provides it | Your organisation must decide it | |
|---|---|---|
| Knowing which data is safe to put into a tool | Yes | No |
| Skill to build, test and run models | Yes | No |
| A written data classification for your content | No | Yes |
| Whether prompts leave your estate for an outside model | No | Yes |
| Retention and access rules for transcripts | No | Yes |
| Who owns the approved knowledge base | No | Yes |
| A tested fallback to a human | No | Yes |
The top two arrive with the people you hire. The bottom five arrive with your governance and your architecture, and well-trained staff cannot make up for them missing.
Architecture is the clearest case. A perfectly trained employee using a system that sends every prompt to an external model has still sent the data. Whether that happens is decided when the system is designed, not when staff are trained. The obligations under the Personal Data Protection Law attach to what the system does. This is not legal advice. Take the specifics of your obligations to qualified counsel.
Governance is the other. SDAIA teaches data governance as a track, which helps, but someone in your organisation still has to write the classification, approve the source documents and own the answers the system gives. A retrieval-based assistant is only as reliable as the approved material behind it, and curating that material is a management responsibility before it is a technical one.
The practical conclusion for buyers is to treat training and architecture as separate items on the plan. Hire and train for operation. Then make sure the design choices the training cannot change, such as where data goes, who reads it and how long it stays, are settled in writing. Our comparison of assistant vendor categories lists the questions that bring those choices into the open, including the ones to ask us.
The honest limits of this article
Every training figure here comes from SDAIA, the Ministry of Education or reporting of their statements. None of it is our measurement, and none of it has been independently audited. The terms differ between announcements: “beneficiaries”, “trained and certified” and “completed” are not defined the same way. We have reported each figure with its own wording and date rather than combining them.
Programme details also move quickly. Cohorts open and close, bootcamp fees and eligibility were not stated in the announcements we read, and the scope of SAMAI 2 was still being described at our publish date. Check Athka X and the official programme pages before you enrol or plan a hiring pipeline around anything written here. We have no part in running any of these programmes.
Sources
- SDAIA Leads Development Drive to Empower National Capabilities in Data and AI (SPA) — Riyadh Daily / Saudi Press Agency, Jun 15, 2026
- A National Celebration: Over One Million Citizens Empowered in Artificial Intelligence Through the "Samai" Initiative — Ministry of Education, Saudi Arabia, Nov 9, 2025
- Saudi Arabia Reaches One Million Citizens Trained in AI Through 'SAMAI' Initiative — Riyadh Daily / Saudi Press Agency, Nov 9, 2025
- Saudi Arabia launches AI initiative to empower 1 million citizens — Gulf Today, Sep 13, 2024
- سدايا تجدد الدعوة للتسجيل في مبادرة مليون سعودي للذكاء الاصطناعي (سماي) — Okaz, Jul 12, 2025
- منتدى أخلاقيات الذكاء الاصطناعي يبحث حوكمة القطاع من الرياض — Al Eqtisadiah, Sep 14, 2026
- Saudi Arabia trains over 1m people in push for ethical AI — Arab News, Sep 15, 2026
- ICAN 2026 concludes; SAMAI 2 launched with 11 ministries (SPA) — Riyadh Daily / Saudi Press Agency, Jan 30, 2026
- Saudi Arabia launches SAMAI 2 workforce AI initiative — Middle East AI News, Jan 29, 2026
- SDAIA launches National Cross-Disciplinary Curriculum in Data and AI for university students (SPA) — Riyadh Daily / Saudi Press Agency, Jan 30, 2026
- Saudi SDAIA brings out 'Athka X' platform at ICAN 2026 — GCC Business News, Jan 30, 2026
- SDAIA Academy launches summer training programmes and bootcamps (SPA) — Riyadh Daily / Saudi Press Agency, Jul 3, 2026
- SDAIA Academy to run 18 data and AI bootcamps under Future Summer initiative — Ajel (English), Jul 21, 2026
- SDAIA Academy opens registration for professional training program in LLMs — Mubasher, Dec 10, 2025
- SDAIA trains 215 students from 28 universities — Middle East AI News, Jun 29, 2026
- SDAIA Academy — Saudipedia, Jan 1, 2026
- Elevate Program — Saudipedia, Jan 1, 2026
- Saudi Arabia to introduce AI curriculum across all public schools starting 2025 — Gulf News, Jul 12, 2025
Common questions
SAMAI, the One Million Saudis in AI initiative, is free, delivered in Arabic and online, according to SDAIA’s registration calls. SDAIA Academy registration is also described as free. Individual bootcamps listed on the Athka X platform did not state fees in the announcements we read, so check each programme page before enrolling.
In June 2026 SDAIA reported more than 1,563,983 beneficiaries across all its training programmes, including 14,495 data and AI specialists and experts, plus more than 278 government leaders. SAMAI specifically passed one million trained and certified citizens in November 2025, and SDAIA put SAMAI completions at more than 1.2 million in September 2026. These are published figures, not independently audited ones.
SAMAI is a national initiative launched in September 2024 by SDAIA with the Ministries of Education and of Human Resources and Social Development. It teaches what AI is, how it is used, its ethics, and practical tools for work and daily life, with a strong focus on responsible use and protecting personal and organisational data. It is AI literacy, not engineering training, and ends with an SDAIA-accredited certificate.
Since January 2026, SDAIA Academy programmes and bootcamps have been listed on Athka X, SDAIA’s national data and AI capacity-building platform, where you register, see schedules and keep a national skills record. Registration opens per cohort, so check the platform regularly. Tracks in 2026 ranged from prompt engineering and responsible AI use to data engineering, computer vision, AI agents and data governance.
It changes the handover more than the architecture. Locally trained staff make it realistic to run and maintain a system after the integrator leaves, with less reliance on imported operators. But training does not decide where prompts and data go, how long transcripts are kept, who owns the approved content, or how the system falls back to a human. Those are design and governance decisions.
Keep reading
LEAP 2026: The Billions in Tech Commitments, and What Has Actually Been Built
LEAP 2026 closed with nearly $15bn in announced investments and agreements. We separate new money from restated figures, valued-at numbers from capital, and plans from services you can use today — with a five-question guide to reading any LEAP headline.
Read Article
Saudi Arabia’s Year of AI 2026: What HUMAIN, Hexagon and a Million Trained Citizens Actually Change for Enterprise Buyers
Saudi Arabia designated 2026 the Year of Artificial Intelligence. The compute, the capital and the training were already built. Here is which of it changes an enterprise AI decision, and what a national programme still cannot answer for you.
Read Article
AI Ethics in Saudi Arabia: What the UNESCO Global Forum in Riyadh Means for Organisations
Riyadh hosted UNESCO’s fourth Global Forum on the Ethics of AI in September 2026. Here is what was verifiably launched, what UNESCO’s and SDAIA’s frameworks ask, and how to turn each principle into a concrete design control for an AI assistant.
Read Article