The unequal AI dividend: Who gets the benefits?

Artificial Intelligence is a global investment phenomenon. Governments fund national AI strategies. Corporations spend billions on AI. Investors consider AI’s potential impact on GDP. But the excitement of the AI moment draws less attention to who gets the AI dividend.
Two dominant foci dominate the discussion: how much capital corporations allocate to AI and how many jobs AI displaces. But far fewer are interested in knowing the economic value AI has already created. Does AI create value by improving productivity? Does it enhance healthcare and education? Does it empower small businesses to do more with less? Do these benefits extend to all countries? The evidence suggests not.
The World Bank reports that 87% of AI models, 86% of AI startups, and 91% of VC funding in AI are in high-income countries, even though they account for only 17% of the world’s population. They also have far superior data centers, computing infrastructure, digital skills, and research.
However, an interesting divergence exists between AI investment and AI adoption. In 2026, the Stanford AI Index recorded global generative AI adoption at 53%. Generative AI adoption was 64% and 61% in the UAE and Singapore, respectively, which is disproportionately high for their respective economic sizes. While the US leads in global AI investment and model building, it ranks at 28.3% on this metric. This highlights an important distinction. AI building and AI model application are related, but distinct economic activities.
The next phase of competition will not be about who develops the largest model, but who can use AI to enhance productivity in their economy. It will not be about creating the next frontier model. A country can see major gains if government agencies and businesses use AI effectively in healthcare, logistics, finance, and the services sector, among others.
This is especially applicable to developing/new economies. Rather than fearing AI will displace jobs in these economies, they can use AI to offset the short-term expertise gap. AI can assist small businesses with customer service and back-office work like accounting. AI can assist farmers in the field, help healthcare professionals manage large amounts of information, and provide support for decision-making. Similarly, it can help government agencies process and analyse large volumes of information.
The World Bank reports on the development of 'Small AI' technology that can solve many problems. However, developing countries face major challenges in connectivity, computing, data, and digital skills required to adopt AI effectively.
AI is also changing the world of work, though the extent and impact on employment remain unclear. Younger software developers are among the first to experience weaker employment prospects, yet evidence to date does not suggest that AI is causing mass unemployment across economies. However, focusing only on employment destruction creates a major blind spot: how are AI-generated productivity gains shared globally? AI may allow one worker to produce twice as much, and that may result in lower-priced products, higher worker wages, higher corporate profits, new products, or job loss. But the technology itself does not determine which outcome prevails. Rather, firm strategy, market concentration, bargaining power, ownership structures, labor institutions, and public policy shape who ultimately receives the AI dividend.
For governments and policymaking worldwide, this matters. The goal is not to measure how much governments invest in AI. It should be to measure how much value the investment creates, such as public services, greater productivity, improved healthcare, and greater economic growth. The global AI race has begun again. The first phase was about building AI technology. The next phase will be using machine intelligence to gain economic advantage. The countries that invest most in AI technology may not win this race. Countries that learn to implement AI technology effectively and ensure the return is widely shared may win.