AI boom needs $6 trillion annual revenue by 2031 to justify infrastructure investments: Report

New Delhi: The rapid buildout of artificial intelligence (AI) infrastructure will require the creation of an estimated $6 trillion in annual revenue by 2031 to justify the capital being deployed with productivity gains alone unlikely to support the economics of the investment cycle, a report said on Tuesday.

According to Bain & Company’s analysis, existing AI applications across consumer and enterprise segments could generate between $1.2 trillion and $1.8 trillion in revenues, leaving a gap of about $4.2 trillion that will need to come from new categories of innovation.

It also identified four major areas for making money with AI like search and advertising, self-driving vehicles and drones, robotics and digital twins and new fields -- medicine, mental health and energy

“Access to AI is no longer the differentiator, absorption is,” said Gurpiar Sibia, Partner and India Head of AI, Insights and Solutions Practice at Bain & Company.

He also said that while Indian firms have access to the same frontier AI models as global peers, competitive advantage will depend on how quickly companies redesign workflows, modernise data systems and deploy AI at scale.

David Crawford, chairman of Bain’s global technology practice, said the industry’s focus on employee productivity was too narrow.

“The economics of AI infrastructure demand trillions in new revenue beyond productivity gains,” he said. Sustainable funding of AI investments may require adding about 1 per cent to annual global GDP growth, according to him.

According to the report, surging demand for AI computing has revived the hardware industry with hardware and semiconductor stocks growing at a 24 per cent compound annual rate between 2020 and 2026, compared with 6 per cent for software stocks.

Custom chips, high-bandwidth memory and advanced packaging are emerging as key growth segments, it said.

However, it flagged that AI is reshaping cybersecurity risks, cutting the time required for a typical cyberattack from about four weeks to roughly 18 hours.

ABOUT THE AUTHOR

IANS

[email protected]