Hyperscalers Face $300 Billion Revenue Hurdle for AI Break-Even

Goldman Sachs estimates that major US AI hyperscalers need to generate $300 billion in annual AI revenue to break even on their heavy spending. The analyst's note highlights that these companies are on track to spend $800 billion on capital expenditures in 2026, with a projected rise to $1.1 trillion by 2027.

The hyperscalers' cloud revenues have accelerated sharply this year, reaching $70 billion above the pre-AI trend in Q2 2026. The announced revenue backlogs exceed $1.5 trillion, indicating significant growth potential.

Goldman's baseline for 2027 spending is that it will beat consensus expectations, but with slower growth and reduced upside surprises. To achieve solid returns, AI users would need to spend about $1 trillion a year on AI applications.