Overview
The article titled "Will worries about AI Doom end the AI Boom? BCA answers", authored by Pranav Kashyap and published on 20‑09‑2026, summarizes a BCA Research report authored by Chief Economist Peter Berezin. The report cautions that the current enthusiasm for artificial‑intelligence (AI) investments may be concealing significant long‑term monetisation challenges and an overstatement of corporate profit margins.
Valuation Context
Berezin notes that the S&P 500 is trading at roughly 19 times forward earnings, which aligns with its ten‑year historical average. This valuation is underpinned by record‑high forward profit margins of 16.7%. Should these margins revert to 2019‑level levels, the index’s forward price‑to‑earnings multiple would rise to approximately 26.7 times, indicating a much higher valuation relative to earnings.
Drivers of Current Margins
The elevated margins are attributed to two primary factors: (1) a decline in real unit labour costs and (2) massive capital spending by the leading hyperscalers—Microsoft, Amazon, Alphabet, Meta, and Oracle. When these technology giants acquire hardware, the outlay is recorded as capital expenditure and depreciated over the asset’s useful life, rather than being expensed immediately, which artificially inflates current profit margins.
Capital Expenditure and Depreciation Outlook
BCA projects that annual depreciation charges for the major hyperscalers will more than double, rising from $255 billion in 2026 to $581 billion by 2029 as total capex reaches $1.16 trillion. Including off‑balance‑sheet spending, total hyperscaler capital expenditure could approach $1.4 trillion per year by the end of the decade.
Revenue Requirements for Sustainable Returns
To generate a realistic 15% pre‑tax return on invested capital and achieve 30% EBITDA margins—figures BCA deems more plausible than Wall Street’s optimistic 50% EBITDA margin forecast—hyperscalers would need to produce $7.4 trillion in annual revenue. When broader AI spending is considered—including non‑hyperscaler activity across China, emerging “neocloud” providers, and private ventures such as SpaceX—the global AI industry would require approximately $10 trillion in annual sales to justify the ongoing capital outlays.
Scale of the Revenue Gap
Berezin highlights that the $10 trillion revenue target is comparable to the total worldwide annual expenditure on either healthcare or food, underscoring the massive scale of the commercial hurdle facing the AI sector.
Conclusion
The BCA report concludes that without achieving the projected revenue levels, the AI boom could face a slowdown as investors reassess the sustainability of current valuations and the true profitability of massive data‑center investments.