LATEST
All analysis →
A data centre on the steppe at night, with wind turbines and lines of light rising into the sky
Home / AI

AI and Mongolia: what changes, and how to adopt it

AI is moving money, jobs and power use worldwide. Mongolia starts from a strong e-government base and a weak AI economy. What other countries did, what it would cost, and where to start.

At a glance

$582bn
Record corporate AI investment in 2025.
Stanford AI Index 2026
0.48/1
Mongolia on the IMF AI Preparedness Index, level with Vietnam. Singapore: 0.80.
IMF AIPI, 2023 data
46th
of 193 on UN e-government in 2024, up from 74th. A strong base to build on.
UN EGDI 2024

Stanford HAI, AI Index 2026, ch. 4 Economy (full chapter PDF linked from cited page) · · desk ai_kpi_invest; IMF DataMapper API, AI_PI · retrieved · desk ai_kpi_aipi; UN E-Government Knowledge Base, Mongolia · · desk ai_kpi_egdi · mongolia-unspun.mn

The world

What AI is doing to economies

Each measure with its caveat, because forecasts differ widely.

Share of jobs exposed to AI

Almost 40% of global employment; about 60% in advanced economies, 40% in emerging markets, 26% in low-income countries. In advanced economies roughly half of exposed jobs may benefit (higher productivity), the rest may see lower labour demand.

IMF, 2024. Caveat: Exposure is not job loss: it counts jobs whose tasks AI could affect, complementing or replacing. Built from occupation-task mapping; says nothing about timing.

Share of jobs exposed to generative AI (narrower measure)

About 25% of global employment in occupations with some GenAI exposure; 3.3% in the highest-exposure category. 34% of jobs in high-income countries vs 11% in low-income. Women more exposed (10.4% of female employment in the top two gradients vs 5.5% for men).

ILO, 2025. Caveat: Lower than IMF because it covers generative AI only and uses a refined task-level index. Again exposure, not displacement; ILO stresses most exposure means transformation of tasks.

Long-run GDP and productivity impact of generative AI

Could raise global GDP by about 7% (~USD 7 trillion) over 10 years and lift productivity growth by 1.5 percentage points a year; ~300 million full-time job equivalents exposed to automation.

Goldman Sachs Research, 2023. Caveat: Older (March/April 2023) scenario estimate from an investment bank, conditional on wide adoption; Goldman itself says most jobs are partially exposed and more likely complemented than replaced. Treat as an upper-range scenario, not a forecast.

Measured macro effect so far

AI-related technology investment added an estimated 0.5 percentage point to US GDP growth in 2025. IMF's April 2026 reference forecast (global growth 3.1% in 2026, 3.2% in 2027) does not include direct productivity gains from AI because adoption is still modest in many sectors. IMF flags downside risk: a sharp repricing if AI profit expectations prove too optimistic, with debt-financed AI investment a concern.

IMF, 2026. Caveat: The 0.5 pp is an investment (spending) effect in one country, not a productivity effect. Productivity gains remain largely unmeasured at macro level.

Productivity and wages in AI-exposed industries

Revenue per employee grew 27% in the most AI-exposed industries vs 9% in the least exposed (2018-2024); AI-skilled workers earned a 56% wage premium in 2024 (25% a year earlier). Based on close to 1 billion job ads.

PwC 2025 Global AI Jobs Barometer, 2025. Caveat: Consultancy study; correlational, not causal (AI-exposed sectors may differ for other reasons). 'Productivity' here is revenue per employee.

Firm adoption — official statistics

EU: 20.0% of enterprises with 10+ employees used AI in 2025, up from 13.5% in 2024 (Denmark 42.0%, Romania 5.2%). US: about 18% of firms used AI (Dec 2025, Census BTOS), over 20% expected in H1 2026.

Eurostat; US Census Bureau BTOS via Federal Reserve FEDS Note, 2026. Caveat: Best comparable measures of broad adoption. BTOS methodology was revised in Nov 2025, so earlier series are not strictly comparable.

IMF Staff Discussion Note SDN/2024/001 · · desk ai_world_1; ILO/NASK, Generative AI and Jobs: 2025 update (research brief) · · desk ai_world_2; Goldman Sachs, Generative AI could raise global GDP by 7% (Apr 5, 2023) · · desk ai_world_3; IMF Annual Report 2026, In focus: AI deployment and disruption · · desk ai_world_4; PR Newswire, PwC 2025 Global AI Jobs Barometer (Jun 03, 2025) · · desk ai_world_5; Eurostat news, 20% of EU enterprises use AI technologies (11 Dec 2025) · · desk ai_world_8 · mongolia-unspun.mn

Other countries

How others are adopting AI

Pick a country: what it did, what it got, and the lesson for Mongolia.

IMF AI Preparedness Index, 0 to 1 ( data)

Hover or tap a bar to see its value and source.

IMF DataMapper, AIPIHigher is more prepared

IMF DataMapper API, AI_PI · retrieved · desk ai_aipi_SGP ai_aipi_EST ai_aipi_KOR ai_aipi_ARE ai_aipi_KAZ ai_aipi_IND ai_aipi_MNG ai_aipi_VNM ai_aipi_RWA; IMF DataMapper API, AI_PI and /groups · retrieved · desk ai_aipi_LIC · mongolia-unspun.mn

Singapore

National AI Strategy 2.0 (NAIS 2.0), Dec 2023; National AI R&D Plan 2025-2030

What they did

Target to more than triple AI practitioners to 15,000 (NAIS 2.0, 2023)

More than SGD 1 billion (~USD 786 m) over 2025-2030 for public AI research centres, applied research and talent (from NRF's SGD 37 bn RIE envelope)

Scaled AI Singapore PhD fellowships and accelerated master's programmes

Lesson for Mongolia

Pair a talent number with a ring-fenced multi-year research budget; small states win on skills and governance, not on scale of compute.

Cointelegraph, Singapore releases National AI Strategy 2.0, plans for 15,000 AI experts · · desk ai_cty1_strategy; Telecom Review Asia, Singapore to Invest More Than SGD 1 Billion in National AI Research Plan Over Five Years · · desk ai_cty1_did1…3 ai_cty1_lesson · mongolia-unspun.mn

Mongolia

Where Mongolia stands

Readiness, on international indices
IMF AI Preparedness Index (AIPI, 0-1)

0.484 (2023 data). Peers: Singapore 0.801, Estonia 0.764, Korea 0.727, UAE 0.628, Kazakhstan 0.552, India 0.493, Vietnam 0.482, Rwanda 0.437; low-income average 0.322. Rank not computed.

UN E-Government Development Index (EGDI)

Rank 46 of 193, EGDI 0.8457 (up from 74th, 0.7209 in 2022). E-Participation rank 37. Telecom Infrastructure 0.9374, Online Services 0.8222, Human Capital 0.7775.

Status today
National Strategy on Big Data and AI

drafted by the Ministry of Digital Development, Innovation and Communications with UNDP; presented May 2025 as a draft for Parliament. Plan: 2025-26 National AI Council, GPU-cluster AI Centre and National Data Repository; 2027 fund 100 research topics, train 300 master's/PhD students, 1,000 'AI ambassadors'; 2030 two international data centres and regional AI exports. No budget published.

No dedicated AI law. Personal Data Protection Law in force since 1 May 2022 is the main binding rule for AI that processes personal data.

E-Mongolia (launched Oct 2020)

1,278 government services online and AI features added with E-Mongolia 5.0 (Sept 2025); 'Digital First' programme approved Sept 2025 targets 90% of public services digital, with projected savings of MNT 1.1 trillion. Earlier count: 1.8 million active users and 1,200 services (May 2024).

World Bank Smart Government II project (USD 40.7 m IDA credit, approved June 2022): online services, cloud and cybersecurity, digital skills for 13,000 people, digitalise 2,000 SMEs.

IMF DataMapper API, AI Preparedness Index (AI_PI) · retrieved · desk ai_ready_2; UN DESA, E-Government Knowledgebase: Mongolia · · desk ai_ready_3; Montsame, draft National Strategy on Big Data and AI presented (20 May 2025) · · desk ai_status_1; Levellers, What is AI regulation in Mongolia? · · desk ai_status_2; Montsame, Digital Transformation: 90 Percent of Public Services to Go Online (10 Nov 2025) · · desk ai_status_3; World Bank press release, Mongolia: New Project Helps the Government Go Digital (6 Jun 2022) · · desk ai_status_9 · mongolia-unspun.mn

What to do

Rough costs are what peers spent, not Mongolian estimates.

1DFI-financed data centre tied to new renewable generation

What peers did: Uzbekistan DataVolt TAS-1: 12 MW, USD 250 m total cost, financed by EBRD, OPEC Fund, Proparco, DEG.

Rough costPeer reference ~USD 20 m per MW of IT capacity (USD 250 m / 12 MW), before any new power plant.
Who actsPrivate developer with EBRD/ADB/IFC; Ministry of Energy for power supply.
Time2-4 years; depends on power.

TokenPost, Uzbekistan Signs More Than $6 Billion in AI Infrastructure Deals (28 Sep 2026) · · desk ai_opt_4 · mongolia-unspun.mn

Updates

AI news that matters for Mongolia,

Each with why it matters. Updated as part of the news radar.

Uzbekistan lines up DFI financing for AI data centres

Reports of over USD 6 bn in data-centre commitments to 2030; DataVolt's 12 MW TAS-1 financed with EBRD, OPEC Fund, Proparco and DEG.

Why it matters: Shows multilateral lenders will fund AI infrastructure in Central Asia.

TokenPost, Uzbekistan Signs More Than $6 Billion in AI Infrastructure Deals (28 Sep 2026) · · desk ai_upd_10 · mongolia-unspun.mn

EU delays high-risk AI Act obligations

Digital Omnibus moves Annex III high-risk obligations from 2 Aug 2026 to 2 Dec 2027 (Annex I to 2 Aug 2028); transparency duties largely remain.

Why it matters: Even the strictest regulator slowed its timetable — evidence that compliance capacity is a real constraint.

Orrick, EU AI Act Update: Digital Omnibus finalizes 8 compliance changes (Jul 2026) · · desk ai_upd_09 · mongolia-unspun.mn

IMF: AI investment supports growth but raises correction risk

April 2026 WEO says AI-related tech investment added ~0.5 pp to US growth in 2025 and warns of sharp repricing if AI profits disappoint.

Why it matters: A tech-equity correction would hit global demand and commodity prices — a channel to Mongolia's copper and coal exports.

IMF Annual Report 2026, AI: Deployment and Disruption · · desk ai_upd_05 · mongolia-unspun.mn

Kazakhstan's AI law takes effect

Central Asia's first AI law, signed Nov 2025, entered into force on 18 Jan 2026.

Why it matters: Mongolia's closest regional peer now has a legal framework, a dedicated AI ministry and a national compute cluster.

Caspian Post, Kazakhstan bets on AI to power its next growth phase · · desk ai_upd_02 · mongolia-unspun.mn

South Korea's AI Basic Act enters into force

Labelling of AI output and duties for high-impact AI took effect on 22 Jan 2026, with a one-year grace period before fines.

Why it matters: First comprehensive AI law in Asia's advanced economies; a template for lighter-touch regulation.

Cooley, South Korea's AI Basic Act: overview and key takeaways (27 Jan 2026) · · desk ai_upd_01 · mongolia-unspun.mn