Goldman Sachs: Tech Giants Need Years of AI Revenue to Break Even

Goldman Sachs has delivered a reality check to AI enthusiasts, suggesting that major tech companies will need years to break even on their investments in artificial intelligence. According to Goldman Sachs strategist Ryan Hammond, hyperscalers like Amazon (AMZN), Oracle (ORCL), and Microsoft (MSFT) must generate around $300 billion in AI revenue over the next few years to recover their costs.

Hyperscalers have seen a surge in cloud revenues this year, with annualized growth at approximately $70 billion above pre-AI trends in Q2 2026. Additionally, announced backlogs for these companies exceed $1.5 trillion. Hammond estimates that AI users would need to spend around $1 trillion annually on AI applications for the hyperscalers to generate solid returns and profit margins.

The stakes are high for investors, who have invested heavily in hyperscaler stocks amid AI optimism. The Roundhill Magnificent Seven ETF (MAGS) has seen an 8% gain over the past month compared to a modest increase in the S&P 500 (^GSPC).