Jim Covello of Goldman Sachs Group Inc. advised investors to prioritize large-scale cloud hyperscalers over semiconductor chipmakers in the current AI trade.

This shift in strategy suggests that the financial winners of the artificial intelligence boom may differ from those of previous technological cycles. By pivoting toward the companies providing the cloud infrastructure, investors are betting on the platforms that manage the AI ecosystem rather than the hardware providers alone.

Covello, who serves as the global head of technology research at Goldman Sachs, said that the current AI infrastructure build-out differs from the earlier Internet era [1, 2]. He said that the present landscape is driven by massive spending from hyperscale cloud providers [1, 2].

In two Goldman Sachs reports examining the AI infrastructure build-out [3], the firm analyzed the incentives driving the market. Covello said that the investment environment has shifted away from the chip-focused growth seen in previous booms.

"We recommend being long the hyperscalers vs. underweight semis," Covello said [1].

The research, which was referenced in reports published in April 2026 and reported by Business Insider in May 2026 [2, 3], highlights a divergence in how capital is flowing into the tech sector. While chipmakers provided the essential tools for the initial surge, the scale of the cloud providers is now viewed as the primary driver of sustainable growth.

Goldman Sachs suggests that the sheer volume of spending by these cloud giants creates a different set of risks and rewards compared to the semiconductor industry. This perspective challenges the prevailing trend of focusing exclusively on the hardware that powers AI models.

“"We recommend being long the hyperscalers vs. underweight semis."”

This guidance indicates a transition in the AI investment thesis from the 'picks and shovels' phase—where chipmakers like Nvidia dominated—to a platform phase. By favoring hyperscalers, Goldman Sachs is signaling that the long-term value may reside in the companies that control the cloud environments where AI is deployed and monetized, rather than those that simply manufacture the silicon.