Goldman's Ryan Hammond: hyperscalers need $300 billion AI revenue to break even

Goldman Sachs strategist Ryan Hammond's math arrives as investors have already renewed their enthusiasm for hyperscaler stocks, with the Roundhill Magnificent Seven ETF up 8% inside of a month against a modest gain for the S&P 500. That contrast is the context for his warning: the market has been pricing in AI optimism even though, by his estimate, the hyperscalers are still far from the roughly $300 billion in AI revenue he says they need to break even.

The body frames the shortfall as a matter of scale and time. Cloud revenues have accelerated sharply this year, annualizing at about $70 billion above the pre-AI trend in the second quarter of 2026, and announced backlogs exceed $1.5 trillion. But Hammond's $1 trillion annual AI-application spending figure for solid returns is a much larger number than what the group currently generates, suggesting the payoff timeline is measured in years rather than quarters.

What the outcome hinges on is whether AI-application spending actually accelerates toward the level Hammond describes. If it does, the hyperscalers could convert their backlogs into the revenue needed for break-even and solid returns; if it does not, the gap between current AI revenue and the $300 billion target is likely to remain wide, and investors who recently bid up these stocks may find the timeline longer than they expected. The body offers no date for when the break-even point might be reached.