The National Artificial Intelligence Roadmap 2021–2025 (AI-Rmap) was published by Malaysia's Ministry of Science, Technology and Innovation (MOSTI) to chart the country's path toward becoming a high-technology, high-income nation by 2030. Built around a "quadruple helix" of government, academia, industry and society (GAIS), it set out six strategies, a national AI governance body, and eleven flagship AI projects.
This page preserves the full roadmap as a navigable reference and pairs it with something the government never published: an honest, four-year accounting of what happened next. The short version — the central governance unit it proposed never materialised as designed, AI policy migrated to a different ministry, and a 2021 plan written before ChatGPT was overtaken by events. The detail is below.
An ambitious blueprint, largely overtaken before its end date.
The AI-Rmap's vision endured, but its central machinery did not. The proposed AI Coordination & Implementation Unit (AI-CIU) was never stood up as designed; in 2023 AI ownership moved from MOSTI to the new Ministry of Digital, and in December 2024 a separate body — the National AI Office (NAIO) — took the helm under a successor 2026–2030 plan. Meanwhile the digital infrastructure the roadmap called for arrived overwhelmingly through private capital, not government programmes.
The AI-Rmap Playbook
The full 102-page roadmap as published by MOSTI in 2021. Read it inline or download the original PDF.
What the roadmap set out to do
The goal. Make Malaysia a high-technology, high-income nation by 2030 by building a "thriving and sustainable AI innovation ecosystem" — one where AI augments jobs, drives competitiveness, and improves wellbeing.
The model. A quadruple-helix partnership (GAIS) of government, academia, industry and society, coordinated through an AI Innovation Ecosystem whose nucleus was the "AI-Catalyst" — a virtual hub hosting consortia that would tackle specific national challenges.
The economics. The roadmap leaned on McKinsey modelling: Malaysia's baseline growth of ~4.4% could gain an additional 1.2 percentage points from AI — framed as a roughly 30% uplift to GDP growth — and targeted lifting R&D spend (GERD) from 1.08% toward 3.5% of GDP by 2030.
It was explicitly a "living document," meant to be revised continuously. In practice, revision came in the form of replacement.
The six strategies
Each strategy as the roadmap defined it, with an honest status based on the best publicly available evidence. Where Malaysia published no progress data, we say so — that absence is itself a finding. Tap to expand.
The centrepiece: an AI Coordination & Implementation Unit (AI-CIU) as the apex government body on AI, with six committees and twenty expert groups, plus a national AI code of ethics and data-sharing guidelines.
- Stand up the AI-CIU + 6 committees + 20 expert groups (Horizon 1, 2021–22)
- National AI Code of Ethics disseminated to all stakeholders
- 100% of ministries/agencies adopting AI cybersecurity policies
What happened: the AI-CIU was never established as designed. AI governance instead moved to the new Ministry of Digital (2023) and then to the National AI Office (Dec 2024) — a different body under a different ministry.
Read Strategy 1 in full → · What happened to the AI-CIU →Intensify fundamental and applied AI research, fund AI Centres of Excellence, and lift national R&D intensity.
- Ring-fence RM10m, then a further RM15m, for AI R&D in priority areas
- Establish 10 AI Centres of Excellence within research institutions
- Move GERD from 1.08% toward 3.5% of GDP by 2030
What happened: no consolidated public reporting against these R&D targets exists. (Larger AI R&D allocations appeared later under Budget 2025 — but as NAIO-era commitments, not roadmap delivery.)
Read Strategy 2 in full →Enable cloud adoption, data-sharing within AI consortia, and nationwide connectivity (JENDELA, 5G) as the substrate for AI.
- 80% of MSMEs (~875,000) adopting cloud computing for AI
- Competitive connectivity costs; 100% broadband access
What happened: the compute substrate arrived — but through a private data-centre boom (Google, Microsoft, AWS, Oracle, Nvidia) worth tens of billions, not the roadmap's government programmes. The MSME-adoption KPI was never publicly reported.
Read Strategy 3 → · The data-centre build-out →Build the AI workforce through school-to-tertiary AI education, an AI Reskilling & Upskilling System (AI-RUS), and talent-attraction schemes.
- 200,000 future AI talents via school programmes
- AI-RUS platform with 500,000 registered employees
- Steep growth targets across data/AI roles (AI Architect +328% by 2025)
What happened: large-scale AI skilling did materialise — but chiefly via later government-industry programmes such as AIForMYFuture (400,000+ Malaysians skilled by mid-2025), not the roadmap's named AI-RUS platform.
Read Strategy 4 →Raise public AI awareness and adoption — social-media engagement, online publications, AI awareness programmes for officials, and an AI SCORE rating for SMEs.
- 12 online publications over five years; podcasts, TV/radio
- AI awareness training for thousands of government officials
- AI SCORE adoption rating across 5,000 SMEs
What happened: public-facing AI literacy continued (e.g. the "AI untuk Rakyat" programme), but the specific roadmap mechanisms and their KPIs were not tracked publicly.
Read Strategy 5 →Operationalise the AI-Catalyst, establish eleven AI Catalyst Consortia to deliver the National AI Use Cases, and form a national AI alliance (MyAI-Alliance).
- AI-Catalyst administration + 11 consortia established in year one
- MyAI-Alliance uniting Malaysia's AI communities
- A self-sustaining AI Innovation Ecosystem
What happened: the AI-Catalyst / consortia model did not become the operating structure. Its role was effectively assumed by NAIO's working groups and the later "AI Malaysia" platform.
Read Strategy 6 → · The AI-IE model →Status key: Superseded replaced by a later structure · Partial happened in altered form · No public data no progress reporting found.
The eleven National AI Use Cases
The roadmap's concrete deliverables — flagship AI projects across five priority areas. Public reporting on their outcomes is scarce; where no result was published, the tracker says so. Filter by sector.
What the 2021 survey found
Before drafting the roadmap, MOSTI ran a national readiness survey. It drew 173 valid responses across the quadruple helix — and revealed a country mostly at the starting line on AI.
Who responded
173 valid responses, by organisation type
State of AI infrastructure & data
Share of organisations at each maturity stage
Most organisations sat at "none" or "initial" — the readiness gap the roadmap set out to close.
The AI talent gap the roadmap projected (headcount growth targets to 2025)
Targeted increase in specialist roles, per the 2021 survey
From AI-Rmap to NAIO: how the plan was overtaken
The roadmap was a five-year plan. Within its own window, the institutions, the technology, and the money all moved past it.
The defining contrast
The roadmap's ambitions ran on a shoestring while the market spent at a different order of magnitude.
Bars are illustrative and not to linear scale — the gap is roughly four orders of magnitude.
Explore the roadmap in detail
Every section of the AI-Rmap, preserved and attributed to MOSTI.
Where Malaysian AI stands now
Independent, continuously-updated coverage of the policy, money and sectors that succeeded the roadmap.
The National Artificial Intelligence Roadmap 2021–2025 (AI-Rmap) is a publication of the Ministry of Science, Technology & Innovation (MOSTI), Putrajaya (ISBN 978-967-19025-5-4). The PDF above is reproduced for reference and research with attribution to MOSTI. The official record is held by MASTIC at mastic.mosti.gov.my.
The four-year review, status assessments and data visualisations on this page are independent analysis by airmap.my, based on the roadmap text and publicly reported developments through 2026. Where Malaysia published no progress data against a target, this is stated rather than estimated. Full sources: airmap.my/sources.