Bain: AI Will Move $4.7 Trillion in Corporate Profits by 2035 — Three Times the Internet's Shift, in Half the Time

Bain & Company projects AI will redistribute $4.7 trillion in global corporate profits between 2025 and 2035 — more than three times the internet’s $1.4 trillion impact, compressed into roughly half the time. The firm attributes the shift to three drivers: productivity gains ($1.1 trillion, 24%), innovation ($2.2 trillion, 47%), and outright market-share redistribution among competitors ($1.3 trillion, 29%). Bain finds AI will structurally transform 71% of the 92 sectors it analyzed, compared with 41% during the internet era — reaching industries the internet largely bypassed, including pharma, healthcare, and manufacturing.

Bain groups the exposure into four clusters — Technology Foundation ($1.5 trillion), Rewired ($1.5 trillion), Augmentation ($1.3 trillion), and Revolution ($0.3 trillion) — and frames AI as a “production technology” that changes how value is created, versus the internet’s role as primarily a distribution technology. Lead author Dunigan O’Keeffe’s warning is the sharper point: “speed matters more than most CEOs realize” — early movers compound an advantage of more data and better-rewired workflows with each deployment, while laggards “cannot purchase their way back” once behind.

That compounding shows up in what companies are already spending, not just what they expect to earn. Enterprise AI spending doubled from $3.5 billion to $8.4 billion in six months even as LLM token prices fell roughly 95% over three years — consumption is outrunning the price collapse, not waiting for it to finish. The same dynamic plays out at the individual-agent level: SaaStr’s own AI marketing agent logged a dense, uninterrupted hour of senior-level work for $13.42, a figure no legally payable US wage undercuts.

For companies still treating AI adoption as a cost-management exercise, Bain’s framing inverts the question. The profit pools aren’t waiting to move — they’re already moving, and the report’s own math says the compounding advantage goes to whoever starts rewiring workflows first, not whoever waits for the safest entry point.