Hyperscalers Must Generate $300 Billion in AI Revenue to Justify Spending, Goldman Sachs Says

The largest U.S. AI hyperscalers need to generate about $300 billion in annual AI revenue in the coming years to break even on their massive infrastructure investments, according to a recent client note from Goldman Sachs. This target highlights the immense scale of capital being deployed to build out artificial intelligence capabilities.

The Scale of AI Investment

Goldman Sachs analyst Ryan Hammond reported that major cloud providers are on track to spend $800 billion on capital expenditures in 2026. Market consensus expects that figure to rise to $1.1 trillion in 2027.

While Goldman's own baseline forecast is for 2027 spending to beat the consensus, the firm also anticipates that the pace of growth and the size of upside surprises will begin to slow. This signals a potential maturation of the initial AI infrastructure build-out phase.

Path to Profitability

To justify this spending, a significant return on investment is necessary. According to the note, hyperscaler cloud revenues have already accelerated, annualizing roughly $70 billion above the pre-AI trend in the second quarter of 2026, and announced revenue backlogs now exceed $1.5 trillion.

Ad

However, for hyperscalers to earn solid returns beyond simply breaking even, Goldman estimates that the total market for AI applications would need to reach about $1 trillion annually. This figure is substantial when compared to the approximately $1.5 trillion spent on all global software this year.

Market Implications

Goldman Sachs acknowledged that many investors are skeptical about the long-term durability of earnings for AI infrastructure stocks, a sentiment that is at least partially reflected in current stock prices. The note pointed out that some semiconductor companies' earnings would have to fall significantly for their stocks to trade at their historical average P/E ratios.

Despite this, the bank stated that hyperscalers also trade at low multiples. It believes that as returns on investment become clearer and spending growth moderates, their shares should find support. Looking ahead, Goldman expects enterprise adoption of AI to become increasingly visible in corporate earnings in the coming quarters, creating both winners and losers in the AI software and services space.