Could AI Doom fears halt the AI boom? BCA weighs in

Artificial intelligence firms may eventually have to bring in $10 trillion in yearly revenue to rationalize the huge surge of capital spending flowing into data centers, BCA Research said in a report.

In that report, BCA Chief Economist Peter Berezin cautioned that investor excitement around AI is concealing significant longer-term challenges in monetization while overstating how strong corporate profit margins really are.

Overstated margins and mounting depreciation costs

The '$10 trillion' revenue shortfall

Under consensus assumptions: with a 15% pre-tax return on invested capital and EBITDA margins of 30%—which BCA considers more plausible than Wall Street's optimistic 50% projection—hyperscalers by themselves would need to produce $7.4 trillion in yearly revenue.

Globally, once non-hyperscaler AI spending is included—covering China, 'neocloud' providers, and private efforts such as SpaceX—the worldwide AI sector would need to generate about $10 trillion in annual sales.