New Delhi: Artificial intelligence (AI) could generate USD 2.52 trillion in annual economic value globally by automating and supporting existing work, but rising deployment costs could significantly reduce these potential gains, according to an October 2026 research report by McKinsey.
The report, titled "The AI value curve: The economics of applying AI to existing work," found that while businesses are rapidly deploying AI, many are yet to determine whether the technology delivers sufficient economic returns.
"Many organisations, however, are still measuring AI adoption instead of value," the report stated.
According to McKinsey, a 30 per cent increase in the effective cost of AI could reduce the estimated global economic value pool from USD 2.52 trillion to USD 1.67 trillion, a decline of around 34 per cent, or USD 850 billion.
In the United States alone, the estimated annual value pool could fall by USD 320 billion, from USD 960 billion to USD 640 billion, under the same scenario.
The report clarified that these estimates represent a modelled scenario rather than a forecast.
The findings come as companies increase investments in AI infrastructure, software and automation, raising questions about whether productivity gains can justify deployment costs.
McKinsey estimated that around USD 7 trillion could be invested in AI infrastructure by 2030, requiring approximately USD 760 billion in annual profit pools to justify the investment.
The research identified financial services as the sector with the largest potential economic gains from AI, estimated at USD 530 billion annually, followed by education at USD 410 billion, technology, media and telecommunications at USD 360 billion, and healthcare at USD 310 billion.
It also found that some of the strongest economic benefits could come from supporting highly skilled, higher-paid professionals rather than simply automating large volumes of lower-wage work.
Occupations involving legal services, healthcare, management and computer-related work showed significant potential for cost-effective AI deployment.
However, the report stressed that the economic benefits of AI vary across countries because of differences in labour costs.
"The AI value curve also shows that AI economics are local, not global," the report stated.
This distinction is particularly relevant for labour-intensive economies, where relatively lower wages can change the financial case for replacing or supporting human tasks with AI.
McKinsey's analysis compares existing labour costs with the full cost of using AI, including technology expenses, human supervision, verification and rework.
The report noted that AI's economic potential depends not just on technological capability, but also on whether companies can deploy it at a sustainable cost while maintaining acceptable work quality.
It suggested that businesses should prioritise applications offering clear economic returns, assess the impact of changing AI costs and plan how to redeploy employees as technology changes existing work.