AI infrastructure costs may far outpace software revenue

Bain & Co’s new report warns that by 2031, the AI industry might need $6 trillion a year just to pay for infrastructure—far more than the $1.2 to $1.8 trillion its current software could generate. That leaves up to a $4.8 trillion shortfall. To bridge the gap, companies are turning to creative financing, like guaranteeing long-term value for chips and data centers, despite razor-thin profit prospects.

• AI could face a $4.8 trillion annual revenue gap by 2031
• Software revenue may not cover massive infrastructure costs
• Vendors are using residual value guarantees for chip/data center debt
• Hardware stocks are growing far faster than software stocks