Bain says AI needs $4.2 trillion in new annual revenue by 2031

Bain says the AI industry needs $4.2 trillion in new annual revenue by 2031 to help fund its computing demand. Bain estimates the industry will need $6 trillion in yearly revenue by then. Existing AI products may bring in up to $1.8 trillion. Bain estimates annual AI infrastructure spending could reach $1.5 trillion by 2031. Bain assumes capital spending will equal about 25% of industry revenue. Consumer AI subscriptions and ads could bring in $200 billion to $400 billion by 2031. Enterprise uses could add $1 trillion to $1.4 trillion. Bain says ads in chatbots and AI-powered search could bring model providers $100 billion to $200 billion or more. Bain identifies autonomous vehicles and industrial automation as a $400 billion opportunity. Bain says physical AI, including robotics and digital twins, could be worth $900 billion. Bain also identifies new applications in drug discovery, mental health support and materials science. Bain estimates capital spending by Microsoft, Google, Amazon, Meta and Oracle could reach $780 billion in 2026. That would be nearly 5 times the level of 3 years earlier. Leading AI data centers are approaching 1 gigawatt of power capacity. Many could approach 2 gigawatts by 2027. Bain expects campuses with 9 gigawatts of capacity by the end of the decade. The report cites Epoch AI figures showing data center size and cost doubling roughly every 12 to 16 months. Bain executive David Crawford said efficiency gains alone would not cover AI infrastructure spending. Crawford said AI infrastructure is being built ahead of demand. Crawford said funding it sustainably would require adding about 1% to the annual global gross domestic product growth rate.