When AI Disruption Arrives Before the Dividend

Uncertain forecasts do not shorten investment lead times. WTO simulations suggest AI could leave global trade 34 to 37 percent higher and GDP 12 to 13 percent higher by 2040 if countries narrow the gaps in digital infrastructure and capability. Daron Acemoglu’s estimate, based on tasks that can currently be measured, implies a much smaller productivity gain over the next decade. These are not estimates of the same thing and should not be averaged. They leave governments with potentially large effects and no settled path of diffusion.

The concentration is less ambiguous. Stanford’s 2026 AI Index reports that private AI investment in the United States reached $285.9 billion in 2025, more than 23 times China’s $12.4 billion. Its census of notable model releases recorded 50 from the United States and 30 from China, compared with five from South Korea and one each from Canada, France, Hong Kong and the United Kingdom. Private-investment data understate Chinese state-backed spending, but the broader pattern is clear: the money and institutions capable of developing advanced AI remain concentrated in two countries.

Labour exposure is more widely distributed. The ILO’s 2025 occupational index finds that one worker in four is in an occupation with some exposure to generative AI, while 3.3 percent of global employment sits in its highest-exposure category. Clerical work remains the most exposed, while professional and technical occupations are becoming more exposed.

Lower measured exposure is not protection. A country can have fewer automatable jobs and still fall behind if firms elsewhere raise productivity faster. It can also lose an export niche without automating much of its own economy. Customer support, back-office processing, software services and routine design work can be displaced through purchasing decisions made by clients abroad.

The World Bank and ILO call this “disruption without dividend”. In many developing economies, workers in automation-vulnerable activities are connected enough to face displacement. Workers whose jobs could become more productive may sit in firms without reliable connectivity, usable data, digital systems or managerial capacity. One group encounters AI through foreign competition before the other encounters it as a useful tool.

The Philippine business-process outsourcing industry was built by connecting a large English-speaking workforce to foreign demand. According to an IMF study of the labour market, the sector generated $35.5 billion in revenue in 2023. It employed about 3 percent of the workforce but produced income equivalent to roughly 7.4 percent of GDP, close to the value of remittances. North America supplied about 70 percent of its market.

Different methods put different boundaries around the exposure. The IMF classifies 36 percent of Philippine employment as highly exposed to AI. The ILO counts 12.7 million workers in occupations with some exposure but only 3.6 percent of employment in the highest category. A World Bank assessment puts roughly one-fifth of workers in highly exposed, low-complementarity jobs and singles out contact-centre work. The numbers are not forecasts of layoffs. Their common finding is that a nationally important export industry contains a large block of tasks that clients can reorganize from abroad.

The industry can still move up the value chain. Philippine firms can use AI to improve quality and retain contracts, and the country has growing higher-value niches in health information, analytics and compliance. The question is whether domestic firms can finance that shift before clients use the same technology to buy less routine work.

The downside channel is direct. A foreign client that replaces contact-centre work with an AI service reduces Philippine export receipts, wages, consumption and payroll-linked revenue. The corresponding profit may accrue to a model provider or cloud platform headquartered elsewhere. Taxes on local digital consumption can recover part of the base. They do not replace a large domestic payroll and the business income distributed around it.

There is not yet strong causal evidence of an AI-driven collapse in Philippine outsourcing. Treating it as inevitable would be irresponsible. Treating it as somebody else’s commercial decision would be equally complacent. The exposure is nationally significant, and the lead time for firm upgrading and new export industries may be longer than the lead time for clients to change suppliers.

One influential American response to mass automation begins with abundance and redistribution. In his 2021 essay Moore’s Law for Everything, Sam Altman proposed an American equity fund capitalized by taxes on large companies and private land, with annual distributions to every adult. The proposal assumes that AI-generated wealth can be captured from appreciating American assets and shared with American citizens.

That assumption is geographical. The companies, their equity and much of their taxable value sit within reach of the state paying the dividend. For a country that mainly consumes AI services, the fiscal arithmetic can run in the opposite direction.

An IMF scenario exercise published in 2026 warns that falling employment and wage income could erode personal-income and payroll tax bases. Higher corporate-tax receipts may only partly offset the loss because capital is mobile, market power is concentrated and countries compete over tax. The problem is sharpest for economies that consume AI services rather than produce them.

The international tax system reinforces the asymmetry. IMF work on cross-border services notes that a remote supplier can earn substantial revenue from a country while leaving it limited rights to tax the associated income under traditional rules. VAT, withholding taxes, digital-services taxes and expanded concepts of taxable presence can recover part of the base. They do not ensure that a country losing a labour-intensive export industry captures the rents earned by the technology replacing it.

Blocking AI to preserve every existing job is not a development strategy. The objective is to shorten the interval between exposure and capability: help firms adopt and adapt the technology, build new sources of income, finance portable skills and keep tax and social-protection systems functioning through the adjustment. Labour policy cannot finance the power, compute, data or company systems on which the new work depends.

Climate and AI are not equivalent crises. Climate policy responds to a physical externality and an agreed need to reduce emissions and strengthen resilience. AI is a privately developed general-purpose technology whose effects can be productive, disruptive or both. The comparison does not justify indiscriminate subsidy, technological nationalism or a public institution with no viable assets to finance.

The useful comparison is the structure of transition. Both changes cut across sectors, depend on complementary infrastructure and capabilities, produce cumulative advantage and distribute costs unevenly. Governments must coordinate long-term investment while the technology and economics are still changing.

Climate finance did not produce a green bank in every country. It produced an architecture: concessional funds for projects that could not carry commercial finance, disclosure rules that made risk visible, common definitions, supervisory work and specialist public financiers able to prepare transactions and absorb selected risks. The functions mattered more than the institutional label.

That experience led to a practical choice. If a government already owned a development bank with capital, clients and operating systems, should it add a specialist mandate or create a new institution? Diana Smallridge and her co-authors answered with a diagnostic in Build or Renovate?: examine the economy and project pipeline, the state’s existing programmes, and the development bank’s mandate, portfolio, staff, governance, financial health and reputation before choosing the institution.

A sound public financier may already possess the balance sheet, market access and client network needed to move quickly. A weak bank with politically directed lending, poor risk management or a conflicting portfolio can be a bad foundation. A new bank carries its own fixed costs, concentration risk and need for a durable pipeline. OECD research on green banks adds the limit that matters for AI: mobilization tools work only when a viable transaction can eventually attract private investors. Where projects, sponsors and markets are too weak, grants and capacity building have to come first.

Before choosing an institution, a government has to separate financeable assets from public obligations. The World Development Report 2026 describes three broad national pathways: adopting existing AI, adapting it to local needs and advancing the frontier. For most developing countries, frontier model development is not a realistic near-term objective. The urgent work is to establish the foundations that let firms and public services use the technology productively.

Compute is the most visible capital requirement. Global data-centre investment reached about $500 billion in 2024, according to the International Energy Agency, while emerging and developing economies outside China held less than 10 percent of global capacity. Yet a national data centre is not automatically a development project. It needs users, power, connectivity and a revenue model. Smaller economies may receive better value from shared regional facilities or contracted cloud capacity.

Firm adoption creates a different financing gap. The OECD’s review of smaller firms finds that connectivity, data, compute, skills and finance operate together. A company may have to clean its records, connect old systems, redesign a process and train staff before an AI tool produces a return. Much of that investment is intangible and difficult to pledge as collateral.

Governments are not converging on a national AI development bank. Their programmes already show four workable architectures.

A government considering an AI-focused public finance institution should answer six questions before drafting legislation.

1. 01

   What national exposure is it responding to?

   Identify the industries, fiscal revenues, regions and worker groups at risk, along with the capabilities that could generate new income. A generic ambition to “lead in AI” is not a mandate.

2. 02

   What position in the AI economy is it financing?

   Broad adoption, local adaptation, data-centre development and frontier research require different systems. The country needs a plausible position grounded in its market, skills, energy system and trading relationships.

3. 03

   Is there a durable pipeline of financeable projects?

   Announced spending is not a pipeline. Projects need sponsors, users, cash flows, infrastructure and credible demand. Public goods and social spending must be separated from assets that can repay capital.

4. 04

   Which failures call for financial instruments?

   Guarantees can address collateral or perceived risk. Equity can support young firms. Leasing can manage hardware risk. Grants can prepare projects or purchase public goods. Regulation and worker protection require other institutions.

5. 05

   Can an existing institution carry the mandate credibly?

   Test governance, technical hiring, risk management, equity and guarantee capacity, reach among firms, impact measurement and conflicts with the existing portfolio. Legal permission alone is a poor test.

6. 06

   Would a new institution improve delivery enough to justify its fixed cost?

   The case is strongest when viable transactions repeatedly fall between mandates, programmes are fragmented and combined technical and financial teams would lower the cost of preparing deals. The institution also needs a way to prove additionality and withdraw when private finance takes over.

A country that passes those tests may have a case for building. The strongest candidates are likely to be large national or regional markets with significant exposure, a defined development pathway, a durable investable pipeline, gaps in risk-bearing or long-term capital and no existing institution able to hold the mandate without serious conflict.

Many countries will not pass. Their transition still needs financing. A ring-fenced window in a development bank, a specialist fund, an innovation mission or a regional facility may carry less institutional theatre and deliver more useful capacity.

Waiting for conclusive labour data may mean waiting until clients, tax flows and investment have already moved. The question is no longer whether public finance has a role. It is which risks merit public capital before the tax base moves with them.