Global Economic Transformation in the AI Era

01 · What I'm tracking

The global story in one page

AI is a general-purpose technology diffusing through a deeply interconnected world economy. Its global impact is governed less by frontier capability than by who can access compute, data, talent, and capital — and by how trade, migration, and governance distribute the gains across nations. The question I keep weighing: convergence or divergence?

Geography

An uneven world map

Capability and value concentrate in a few hubs. The rest of the world participates mainly as adopters, data sources, and markets — unless deliberate strategy enables leapfrogging.

Mechanism

Diffusion across borders

Gains spread via trade, cloud platforms, open models, talent flows, and FDI. Frictions — export controls, data localization, infrastructure gaps — shape who benefits.

Stakes

Power & prosperity

AI reshapes comparative advantage, the offshoring model, and geopolitical leverage. Compute and chips become strategic resources akin to energy.

🎯

What I keep coming back to: The global economy faces a fork. One path is divergence — a widening "AI divide" where compute-rich economies pull ahead and the offshoring ladder is kicked away from developing nations. The other is broad-based convergence — where cheap cognition, open models, and capacity-building let lagging economies leapfrog. Which path wins is a policy and coordination choice, not a technological inevitability.

02 · The Big Picture

Global AI value system — how capability flows across the world

A small set of capability hubs produce frontier models and compute; these flow through global platforms and trade into adopter economies; output and surplus are then distributed across nations — with feedback loops that either concentrate or spread advantage.

FIG 1

Capability hubs → global platforms → adopter economies → distribution

How I read it left→right: a few hubs make compute and models, global platforms and trade carry them outward, adopter economies absorb them at very different depths, and the surplus redistributes into income, trade competitiveness, and power. The teal loop = capacity-building that spreads gains; the red loop = forces that widen the divide.

03 · The World Map

Regional landscape — who's positioned how

Major economies and blocs occupy very different positions on capability, adoption readiness, and strategy. This is the heart of the "global" story: the transformation is not one event but many, at different speeds.

TBL 1

Comparative position of major economies & blocs

Region / Bloc	Frontier capability	Adoption readiness	Key strength	Strategic posture
United States	Leader	High	Frontier labs, capital, compute, talent	Lead capability; export-control compute
China	Near-frontier	High	Scale, data, manufacturing, state push	Self-reliance; open models; applications
EU	Strong follower	Medium	Market size, regulation, industry, talent	Trustworthy AI; regulation-first; sovereignty
UK	Strong follower	High	Research, finance, AI safety leadership	Pro-innovation + safety convening
India	Rising	Medium	Talent pool, IT services, large market, DPI	Adoption at scale; AI for development
Gulf (UAE, KSA)	Rising	Medium	Capital, energy, sovereign compute build-out	Buy capability; become compute hubs
Japan / Korea	Strong follower	High	Hardware, robotics, advanced manufacturing	Physical-AI & industrial integration
Southeast Asia	Adopter	Medium	Young workforce, digital economy growth	Adopt & localize; regional data hubs
Latin America	Adopter	Low–Med	Resources, growing tech ecosystems	Adoption; talent; nearshoring leverage
Africa	Adopter	Low	Demographics, mobile-first, leapfrog upside	Infrastructure & skills; localized AI

Legend: Adopter Rising Follower/Medium High/Near-frontier Leader. Positions are directional, not precise rankings — they move as compute, capital, and policy shift.

FIG 2

Capability vs. adoption readiness — where regions sit

The top-right is the leadership zone (high capability + high diffusion capacity). Most of the world's population sits toward the lower-left — the leapfrog opportunity is moving these economies rightward (adoption) even without frontier capability, by leveraging open models and cheap inference.

04 · The Concentration Problem

The global AI divide

Three scarce inputs — compute, talent, and capital — are highly concentrated. This concentration is the single biggest force pushing the world toward divergence rather than shared prosperity.

FIG 3

Concentration of the three scarce inputs (illustrative shares)

When the inputs concentrate, so does the capability — and absent countervailing forces, so do the gains. The policy challenge is to broaden access to all three without compromising safety or security.

Compute

The new strategic resource

Access to GPUs, fabs, and data-center energy is gated by capital and, increasingly, by export controls. Compute-poor nations risk permanent dependence.

Talent

Brain flows

Elite researchers cluster in a few hubs; migration policy and local universities determine whether other nations build or lose talent.

Capital

Investment gravity

Venture and corporate AI investment concentrates where ecosystems already exist, reinforcing leads through a flywheel effect.

Data

Language & local relevance

Models trained on dominant languages underserve others. Local-language data and sovereign datasets are a leapfrog lever.

Open models

The great equalizer?

Open-weight models let smaller nations deploy capability locally and cheaply — the strongest force against divergence.

Infrastructure

The on-ramp

Electricity, connectivity, cloud regions, and digital public infrastructure determine whether adoption is even possible.

05 · Trade & Value Chains

How AI rewires global value chains & trade

AI changes comparative advantage itself. The labor-cost arbitrage that built the offshoring model weakens when cognition is cheap and automatable — with profound consequences for development pathways.

Disruption

The offshoring ladder wobbles

Services offshoring (call centers, BPO, basic coding/back-office) is highly automatable — a key development on-ramp is at risk.
Reshoring pressure grows as automation erodes the labor-cost rationale for distant production.
"Premature deindustrialization" risk: developing economies may lose the manufacturing/services ladder before climbing it.

Opportunity

New routes to participate

Digitally-delivered services can scale globally with AI leverage — small firms reach world markets.
AI-enabled productivity lets smaller economies punch above their weight in tradable services.
Nearshoring + AI creates regional value-chain hubs (e.g., Latin America, SE Asia, North Africa).
Data & localization services become exportable products.

🔁

The trade paradox: AI both globalizes (any digital service is instantly tradable, language barriers fall) and localizes (production returns to capability hubs; data-localization and sovereignty rules fragment the internet). Expect a more regionalized, bloc-structured global economy rather than simple hyper-globalization.

06 · Adoption Across Borders

Four phases — and why regions move at different speeds

The same S-curve plays out globally, but the gap between leaders and laggards widens in the middle phases before (potentially) narrowing as diffusion broadens.

FIG 4

Divergence then convergence — leader vs. follower adoption

Leaders capture early surplus and reinvest, pulling ahead in phases 2–3. Whether followers catch up (green) or fall permanently behind (red gap) depends on access to compute, open models, skills, and capital.

PHASE 1 · 2022–24

Experimentation

Pilots concentrate in advanced economies and large firms. The Global South watches; access gaps emerge early.

PHASE 2 · 2024–27

Deployment

Leaders productionize agents in core sectors. Compute & export controls harden. The divide widens.

PHASE 3 · 2027–30

Restructuring

Global value chains reshape; offshoring models stressed. Open models & cheap inference open leapfrog windows.

PHASE 4 · 2030–35

Re-globalization

A more regionalized, bloc-structured economy settles. Convergence possible where capacity was built.

07 · Global Labor

Work across a divided world

Exposure differs sharply by a country's economic structure. Service-heavy advanced economies face cognitive automation; manufacturing- and agriculture-heavy developing economies face different timing and a different ladder problem.

Advanced economies

Cognitive disruption, augmentation upside

High white-collar exposure, but also high capacity to augment and redeploy. Aging workforces make productivity gains valuable; strong safety nets cushion transition.

Emerging economies

Mixed exposure, real upside

Large young, educated, English-capable workforces (e.g., India, Philippines) face BPO/services automation — but also have the most to gain from AI-leveraged services exports.

Developing economies

Slower hit, ladder risk

Lower near-term exposure (more physical/informal work), but the bigger danger is the development ladder being pulled away before they climb it.

FIG 5

The global skills barbell + the demographic angle

For young, fast-growing populations (much of Africa, South Asia, parts of SE Asia), AI is opportunity or threat depending entirely on skills and access. The "demographic dividend" turns into a liability if those workers are stuck in the squeezed middle.

08 · Global Growth

What it means for world GDP & convergence

AI can lift global output substantially — but the distribution across nations determines whether the world converges (poorer countries catch up) or diverges (the gap widens).

Channel 1

Capital build-out

Global capex on compute, data centers, chips, and energy is a near-term GDP engine — concentrated where the infrastructure is built.

Channel 2

Productivity diffusion

The main long-run channel: cheaper cognition raises output per worker — if it diffuses beyond the hubs into the global services economy.

Channel 3

Accelerated innovation

If AI speeds up R&D and scientific discovery, it could lift global TFP — a potential rate effect on world growth, not just a level effect.

FIG 6

Two trajectories for the global income gap

This is the defining global question. The technology is identical in both branches; what differs is whether access to compute, models, skills, and capital is broadened or hoarded.

09 · Power & Security

The geopolitics of compute

AI has become a domain of great-power competition. Semiconductors, energy, data, and standards are the new strategic terrain — and a chokepoint-laden global supply chain sits underneath it all.

FIG 7

The AI compute supply chain — chokepoints & leverage

The supply chain's concentration makes it both fragile (single points of failure) and strategically powerful (export controls as statecraft). This is why "compute sovereignty" has become a national priority for many governments.

Export controls

Chips as statecraft

Restrictions on advanced chips/tools are reshaping global tech alliances and accelerating self-reliance drives.

Energy

Power as the new oil

Data-center electricity demand turns energy policy into AI policy; nations with cheap power gain an edge.

Standards

Rule-setting power

Whoever sets safety, data, and interoperability standards shapes global markets (the "Brussels/Beijing/DC effect").

Sovereignty

Sovereign AI

Nations build domestic compute, local-language models, and data residency to reduce dependence.

10 · The Strategy

Playbooks for the global system

Different actors hold different levers. I use this stack to check whether national strategy, multilateral coordination, and firm behavior are pulling toward convergence or widening the divide.

FIG 8

The global strategy stack

Convergence ("Broad Global Prosperity") requires all three layers to pull together: multilateral access & safety, smart national strategy, and firms that diffuse capability rather than hoard it.

For Leading Economies

Sustain the lead, share the gains

Invest in frontier + diffusion: R&D and economy-wide adoption (especially SMEs & public sector).
Secure compute & energy: chips, data centers, clean power, supply-chain resilience.
Reskill & modernize safety nets: manage domestic transition; avoid backlash.
Lead on safety & standards: set credible, interoperable rules.
Manage geopolitics responsibly: balance security controls with avoiding a destabilizing global divide.

For Emerging & Developing Economies

Leapfrog, don't get left behind

Adopt aggressively: use open models + cloud to deploy capability without building frontier labs.
Build foundations: connectivity, electricity, digital public infrastructure, local-language data.
Invest in skills: turn demographic dividends into AI-leveraged workforces.
Find niches: AI-enabled tradable services, nearshoring, sector specializations.
Pool resources regionally: shared compute, data, and standards across neighbors.
Protect the ladder: diversify beyond automatable BPO/manufacturing.

For Multilateral Bodies

Steer toward convergence

Broaden access: compute access programs, open models, and AI capacity-building for the Global South.
Coordinate safety: shared evals, standards, and norms (avoid a race to the bottom).
Govern data & trade: rules for cross-border data, digital trade, and IP.
Fund the transition: development finance for AI infrastructure & skills.
Manage systemic risk: coordination on security, misuse, and stability.

For Global Firms

Operate in a fragmenting world

Redesign workflows globally: deploy AI across the value chain, not just headquarters.
Build for fragmentation: data residency, multi-region, multi-model resilience.
Localize: language, regulation, and market-specific deployment.
Manage geopolitical risk: diversify compute/vendors; scenario-plan for controls.
Invest in local talent & ecosystems: turn presence into capability.

11 · What Could Go Wrong

Global risk & governance map

The global scale adds cross-border risks that no single nation can manage alone — making coordination both essential and hard.

FIG 9

Global risk quadrant — likelihood × severity

The signature global risks: a widening AI divide, geopolitical conflict over compute, cross-border supply-chain shocks, and global misinformation. Tail risks (systemic safety) demand international coordination precisely because no nation can contain them alone.

Divergence

A two-speed world

Compute-rich nations pull away; developing economies lose the development ladder — destabilizing migration, trade, and politics.

Geopolitics

Tech-bloc fragmentation

A "splinternet" of rival AI ecosystems, export-control escalation, and reduced cooperation on shared risks.

Systemic

Safety & security

Cross-border misuse (cyber, bio, influence ops), model proliferation, and longer-horizon advanced-AI safety.

Resource

Energy & environment

Global compute energy demand strains grids and climate goals; water and materials pressures.

Stability

Labor & social unrest

Rapid transitions without safety nets risk backlash, populism, and disorderly migration.

Governance

The response toolkit

International standards & evals, compute-access programs, development finance, trade rules, and crisis coordination.

12 · Where This Goes

Global scenarios, 2030–2035

Two axes define the global outcome space: how broadly access is shared (concentrated vs. distributed) and how cooperative the world order is (fragmented vs. coordinated).

FIG 10

Scenario matrix — access × world order

My north star is the top-right — Broad Global Prosperity: coordinated safety and rules and widely distributed access. The most dangerous quadrant is the bottom-left — fragmented and hoarded. Section 10 is about what it would take to move the world up and to the right.

13 · Track Progress

Global metrics that matter

Signals I watch to know whether the world is converging or diverging — and how individual nations are positioned.

Global / System level

Convergence vs. divergence signals

Cross-country income gap	Is the rich–poor nation gap widening or narrowing?
Compute distribution	Share of global compute outside top hubs
Open-model adoption	Capability deployed in non-frontier economies
Global productivity growth	Is the world J-curve turning up?
Trade in digital services	Who's exporting AI-enabled services?
Tech-bloc fragmentation	Export controls, data localization, standards splits
Energy & emissions	Data-center power demand & intensity

National level

Country readiness dashboard

AI readiness index	Infrastructure, skills, governance, innovation
Compute access	Domestic + accessible cloud capacity
Talent pipeline	Graduates, retention, net migration of skills
Adoption rate	Firm & public-sector deployment depth
Digital infrastructure	Connectivity, electricity, DPI coverage
Local-language capability	Models & data for the national language(s)
Transition support	Reskilling spend, safety-net coverage

⚠️

Leading indicator I watch: the gap between capability and adoption readiness within and across countries is the single best early signal of which scenario the world is heading toward. A widening gap = divergence; a closing gap = convergence.

14 · Remember This

Ten things I keep on my desk

Not commandments — reminders I re-read when the global noise gets loud.

1 · It's a global GPT, diffusing unevenly

Impact depends on cross-border access to compute, data, talent, and capital — not frontier capability alone.

2 · Convergence vs. divergence is THE question

Same technology, opposite outcomes. Policy and coordination decide the branch.

3 · The inputs are dangerously concentrated

Compute, talent, and capital cluster in a few hubs — the core force toward divergence.

4 · Compute is the new strategic resource

Chips and energy are geopolitical leverage; "compute sovereignty" is now national strategy.

5 · The offshoring ladder is at risk

Cheap, automatable cognition threatens the development on-ramp for emerging economies.

6 · Open models are the great equalizer

The strongest force against divergence — they enable leapfrogging without frontier labs.

7 · Expect regionalization, not hyper-globalization

AI both globalizes and localizes; a bloc-structured economy is the likely result.

8 · Demographics cut both ways

Young populations are a dividend only if workers reach the AI-leveraged end of the barbell.

9 · Tail risks need international coordination

Misuse, safety, and supply shocks cross borders; no nation can contain them alone.

10 · Aim for Broad Global Prosperity

Coordinated rules plus distributed access — the only quadrant that's high-growth, equitable, and stable.

15 · References & Sources

Where the ideas in this note come from

Annotated bibliography behind the convergence-vs-divergence thesis, global value-system map, regional positioning, AI-divide concentration chart, trade & GVC disruption, cross-border adoption curves, global labor barbell, growth trajectories, compute geopolitics, strategy stack, scenario matrix, and KPI dashboard. Section tags (e.g. §05) show where each source is used. Diagrams and operating rhythms are my synthesis unless noted.

Scope. Synthesis of global macro, development economics, trade, and AI-industry sources (May 2026). Hero-strip figures (e.g. $15–20T GDP range, ~70% value concentration, three blocs) blend McKinsey, Goldman Sachs, PwC, IMF, and Stanford AI Index estimates — directional ranges, not forecasts for any country or firm. Not legal, financial, trade, or development advice.

Citations are numbered continuously [1]–[n] within this section.

Global GPT diffusion, convergence & the productivity J-curve (§01, §06, §08)

Bresnahan, T. F. & Trajtenberg, M., “General Purpose Technologies: ‘Engines of Growth’?” Journal of Econometrics, 1995. GPT framework for uneven cross-border diffusion — §01 “general-purpose technology.” doi.org/10.1016/0304-4076(94)01598-T — §01.
Brynjolfsson, E., Rock, D., & Syverson, C., “The Productivity J-Curve: How Intangibles Complement General Purpose Technologies.” American Economic Journal: Macroeconomics, 2021. Leaders adopt before measured productivity rises — background for §06 leader/follower curves and §13 global productivity KPI. doi.org/10.1257/mac.20180386 — §06, §13.
David, P. A., “The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox.” American Economic Review (Papers & Proceedings), 1990. Long lags before GPT gains appear in statistics — supports §06 widening-gap phase. doi.org/10.1257/aer.90.2.355 — §06.
IMF, Gen-AI: Artificial Intelligence and the Future of Work (Staff Discussion Note SDN/2024/001). January 2024. ~40% of global jobs AI-exposed; higher exposure in advanced economies; inequality channels — §01 fork and §07 labor cards. imf.org — §01, §07, hero.
IMF, Exposure to Artificial Intelligence and Occupational Mobility: A Cross-Country Analysis. Working Paper, June 2024. Cross-country AI exposure and mobility — TBL 1 regional positioning and §07 exposure differences. imf.org — §03, §07.

Macro estimates & global GDP potential (§08, hero stats, FIG 6)

Goldman Sachs Global Investment Research, “Generative AI Could Raise Global GDP by 7%.” 2023. ~$7T cumulative boost over a decade — lower bound of hero $15–20T range when combined with longer horizons. goldmansachs.com — hero, §08.
McKinsey Global Institute, The Economic Potential of Generative AI: The Next Productivity Frontier. June 2023. $2.6–4.4T annual value across use cases; automation of 60–70% of employee time in analyzed activities — hero $2–4T adjacent range and §08 productivity channel. mckinsey.com — hero, §08.
PwC, Sizing the Prize: Global Artificial Intelligence Study. 2017. Up to $15.7T global GDP contribution by 2030 — upper anchor for hero cumulative GDP range. pwc.com — hero, §08.
IMF, Broadening the Gains from Generative AI: The Role of Fiscal Policies (Staff Discussion Note, June 2024). Policies to spread gen-AI gains across income groups and regions — §10 multilateral “steer toward convergence” bullets. imf.org — §08, §10, §12.
Acemoglu, D., “The Simple Macroeconomics of AI.” NBER Working Paper 32487, 2024; Economic Policy, 2025. Task-based upper bound on TFP gains — counterweight to bullish macro scenarios in FIG 6. nber.org/papers/w32487 — §08.
Aghion, P., Jones, B. F., & Jones, C. I., “Artificial Intelligence and Economic Growth.” NBER Working Paper 23928, 2017. AI accelerating idea production — §08 Channel 3 “rate effect.” nber.org/papers/w23928 — §08.

Regional capability, the AI divide & concentration (§03–§04, FIG 2–3, hero ~70%)

Stanford HAI, 2025 AI Index Report — Economy chapter. 2025. Private AI investment, compute, publications, and adoption by country — backbone for TBL 1, FIG 2, FIG 3 bars, and hero ~70% / ~85% concentration stats. hai.stanford.edu/ai-index — hero, §03–§04, §13.
OECD, AI in Work, Innovation, Productivity and Skills (AI-WIPS) programme & Employment Outlook AI chapters. 2023–25. Cross-country adoption readiness and labor exposure — §03 adoption-readiness axis and §13 national KPIs. oecd.org/ai — §03, §13.
Centre for the Governance of AI (GovAI), State of AI Policy & national capability tracking. Compute access, model development, and policy by nation — TBL 1 posture column. governance.ai — §03, §09.
Center for Security and Emerging Technology (CSET), AI Compute Supply Chain analyses. Geographic concentration of advanced compute and fabs — §04 compute card and §09 FIG 7. cset.georgetown.edu — §04, §09.
International Energy Agency, Energy and AI (special report). 2025. Data-center electricity demand and regional concentration — §04 infrastructure card and §09 energy card. iea.org/reports/energy-and-ai — §04, §09, §13.

Trade, global value chains & the development ladder (§05, §07)

Rodrik, D., “Premature Deindustrialization.” Journal of Economic Growth, 2016. Developing economies peaking industrialization at lower income levels — §05 “premature deindustrialization” and offshoring-ladder risk. doi.org/10.1007/s10887-015-9122-3 — §05, §14.
Rodrik, D., “An African Growth Miracle?” Journal of African Economies, 2018. Services-led development pathways and the limits of manufacturing-led catch-up — §05 opportunity/disruption cards. doi.org/10.1093/jae/ejx030 — §05.
International Labour Organization, Generative AI and Jobs: A Global Analysis of Potential Effects on Job Quantity and Quality (Working Paper 96). 2024. Global augmentation vs. automation; clerical/BPO exposure — §05 services offshoring and §07 BPO barbell. ilo.org — §05, §07, §14.
Eloundou, T., Manning, S., Mishkin, P., & Rock, D., “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models.” arXiv, 2023. Global task exposure framework originating in US O*NET — §07 routine cognitive / BPO exposure. arxiv.org/abs/2303.10130 — §07.
World Trade Organization, World Trade Report 2023: Re-globalization or Fragmentation? 2023. Digital trade, services, and bloc formation — §05 trade paradox and §12 regionalization scenario. wto.org/wtr23 — §05, §12.
McKinsey Global Institute, Global Flows: The Ties That Bind in an Interconnected World. 2024–25 updates on trade, FDI, and digital flows — §02 “Trade & FDI” platform node and §05 nearshoring opportunities. mckinsey.com/mgi — §02, §05.

Task framework, labor & demographics (§07, FIG 5)

Acemoglu, D. & Restrepo, P., “Automation and New Tasks: How Technology Displaces and Reinstates Labor.” Journal of Economic Perspectives, 2019. Displacement vs. reinstatement across tasks — §07 three-economy-type cards. doi.org/10.1257/jep.33.2.3 — §07.
World Economic Forum, The Future of Jobs Report 2025. 2025. Global skill churn, AI-related roles, and reskilling needs — §07 AI-leveraged end of barbell and §10 skills bullets. weforum.org — §07, §10, §13.
World Bank, World Development Report 2024: The Middle-Income Trap. 2024. Development ladder, demographic dividends, and productivity transitions — §07 demographic angle in FIG 5. worldbank.org/wdr2024 — §07, §14.
United Nations Department of Economic and Social Affairs, World Population Prospects 2024. Regional demographic profiles — §03 Africa/India rows and §07 dividend discussion. population.un.org/wpp — §03, §07.

Open models, leapfrogging & digital public infrastructure (§02, §04, §06, §10)

Meta AI, Llama open-weight model releases & responsible-use guides. 2023–25. Widely deployed open weights enabling local inference — §04 “great equalizer” card and §06 convergence arrow. ai.meta.com/llama — §04, §06, §14.
Hugging Face, Open LLM Leaderboard & ecosystem reports. Benchmarking and diffusion of open models globally — §04 open-models card and §13 open-model adoption KPI. huggingface.co — §04, §13.
India Stack / digital public infrastructure literature (NPCI, World Bank DPI reports). Large-scale public digital rails as adoption accelerators — §03 India row and §10 “build foundations” for emerging economies. worldbank.org/dpi — §03, §10.
DeepSeek AI & subsequent industry analyses on efficient frontier models (2025). Lower compute paths to strong capability — §06 Phase 3 leapfrog window (verify latest technical reports). — §06.

Compute geopolitics, export controls & sovereignty (§09, FIG 7)

Miller, C., Chip War: The Fight for the World’s Most Critical Technology. Scribner, 2022. Semiconductor concentration and statecraft — §09 compute supply chain and §04 strategic resource framing. — §04, §09.
CSIS, “A Seismic Shift: The New U.S. Semiconductor Export Controls.” 2022. October 2022 BIS rules as technology-denial statecraft — §06 Phase 2 divide-widening and §09 export-controls card. csis.org — §06, §09.
Farrell, H. & Newman, A. L., “Weaponized Interdependence: How Global Economic Networks Shape State Coercion.” International Security, 2019. Chokepoint leverage via global networks — §02 red “AI divide” loop and §05 trade paradox. doi.org/10.1162/isec_a_00351 — §02, §05, §09.
European Parliament & Council, Regulation (EU) 2024/1689 (AI Act). 2024. Brussels effect on global standards — §09 standards card (“Brussels/Beijing/DC effect”). eur-lex.europa.eu — §09, §10.

Global governance, safety & multilateral coordination (§10–§11)

United Nations, Governing AI for Humanity — High-level Advisory Body final report. September 2024. Global governance gaps and proposed scientific panel — §10 multilateral layer. un.org/governing-ai-for-humanity — §10, §11.
International AI Safety Report 2025 (Y. Bengio, chair). January 2025. Shared scientific assessment across governments — §11 systemic-risk card. internationalaisafetyreport.org — §11.
OECD, OECD AI Principles. 2019 (updated 2024). Interoperable trust and accountability norms — §10 “lead on standards” and §11 governance toolkit. oecd.org/ai-principles — §10, §11.
NIST, Artificial Intelligence Risk Management Framework (AI RMF 1.0). U.S. Department of Commerce, 2023. Risk-based governance template adopted globally — §10 firm/global operating models. nist.gov/ai-rmf — §10, §11.
World Bank & IMF development-finance AI initiatives (Digital Development Partnership, AI capacity-building discussions, 2024–25). Infrastructure and skills finance for lower-income economies — §10 “fund the transition.” worldbank.org/digitaldevelopment — §10.

Inequality, labor share & distribution across nations (§08, §11–§12)

Karabarbounis, L. & Neiman, B., “The Global Decline of the Labor Share.” Quarterly Journal of Economics, 2014. Falling labor share since the 1980s — §08 distribution channel and §12 Managed Oligopoly quadrant. doi.org/10.1093/qje/qju032 — §08, §12.
IMF, analyses of AI and the labor income share (follow-on to SDN/2024/001). Cross-border inequality if AI complements capital and high-skill labor — §01 divergence fork and FIG 6 red path. imf.org/digital/economy — §01, §08.
Autor, D. H. et al., “The Fall of the Labor Share and the Rise of Superstar Firms.” Quarterly Journal of Economics, 2020. Global platform concentration — §04 capital flywheel and §11 tech-bloc fragmentation. doi.org/10.1093/qje/qjaa006 — §04, §11.

Author synthesis & companion notes

Truong, L., Global Economic Transformation in the AI Era — personal working notes. May 2026. Original diagrams (FIG 1–10, TBL 1), global strategy stack, scenario matrix, risk quadrant, and ten desk principles. LinhTruong.com — all sections.
Truong, L., companion notes: Economic Transformation in the AI Era and Geopolitics Transformation in the AI Era. Firm-level and great-power frames cross-referenced in §02 loops, §09 compute geopolitics, and §12 Broad Global Prosperity aim. Same author collection. — §02, §09, §12.

📎

Before you quote externally: FIG 3 concentration bars are illustrative — verify against the latest Stanford AI Index and OECD investment data before citing specific shares. The hero $15–20T GDP range blends PwC (2030), McKinsey (annual use-case value), and Goldman (10-year cumulative) estimates on different bases; do not treat them as interchangeable. Regional positions in TBL 1/FIG 2 are directional snapshots that shift with export controls, open-model releases, and national AI strategies.