Goldman Sachs Warns AI Chip Rally Is 'Unsustainable,' Points to Hyperscalers as Next Winners

Chip Stocks Led the First AI Wave, but Goldman Sees a Rotation Ahead

AI-linked semiconductor names have been the defining trade of the early artificial intelligence boom, but Goldman Sachs is urging investors to look beyond chipmakers as the market searches for the next phase of winners.

In a Monday note, Goldman’s analysts said the AI rally has become increasingly concentrated in infrastructure-related stocks, particularly AI chipmakers, even as investors begin repositioning for what comes next across the broader AI value chain.

'Unprecedented and Unsustainable' Market Dynamic

Goldman’s James Covello, head of global equity research, argued that the current relationship between chip suppliers and their customers has become unusually distorted.

“The general idea is that chip companies are supposed to thrive when their customers thrive,” Covello wrote. “They are not supposed to be thriving at the expense of the companies higher in the chain.”

Covello said today’s setup—where semiconductor firms are reporting record revenue and profits while other parts of the AI ecosystem continue spending heavily—appears “unprecedented and unsustainable.”

Nvidia’s Run Highlights the Infrastructure-Led Rally

The shift in Goldman’s view follows a multi-year surge in AI-exposed chip stocks, a move led by Nvidia, which has emerged as one of Wall Street’s standout winners. The rally has been fueled by aggressive capital spending from major technology companies racing to build AI infrastructure, including large-scale data centers.

However, Goldman noted that a key question is becoming more central to investors: which companies will ultimately convert massive AI outlays into durable, long-term earnings growth.

Goldman: Hyperscalers Could Outperform Semiconductor Names

Goldman now expects hyperscalers—the largest cloud and compute platforms—to outperform semiconductor companies from here, highlighting a potential rotation away from the most crowded infrastructure trades.

The bank said investors have grown cautious about the returns hyperscalers can generate from their AI spending. But that skepticism could shift if evidence emerges that enterprise customers are realizing benefits.

“Investors have become fairly skeptical of the returns the hyperscalers are likely to deliver,” Goldman wrote. “If enterprises begin to show returns from AI spending, investors may be willing to pay higher multiples on these stocks again.”

A Second Tailwind: Spending Discipline and Cash Flow

Goldman also outlined another path for hyperscalers to gain relative strength: a pullback in AI investment designed to improve cash flow.

Under this scenario, hyperscalers could benefit even if enterprise AI adoption remains difficult in the near term. At the same time, a reduction in infrastructure buildouts would likely weigh on semiconductor demand, creating downside pressure for chip stocks while supporting hyperscaler equity performance.

The Key Risk: More Spending Without Clear Enterprise Returns

Goldman flagged a major risk to its thesis: hyperscalers may continue to pour capital into AI infrastructure without seeing meaningful returns from enterprise adoption.

If that happens, the bank said the current dynamic would likely persist, with semiconductor companies continuing to capture a disproportionate share of gains tied to AI spending.

Coinasity's Take

Goldman’s message is less about fading AI and more about where the profit pool ultimately settles. When infrastructure suppliers are thriving while their biggest customers shoulder mounting investment burdens, markets typically start looking for a rebalancing. If enterprise AI begins translating into measurable returns—or hyperscalers turn more disciplined on spending—leadership could shift from AI chipmakers to hyperscalers and deployment-focused firms. The critical variable remains whether AI investment can move from capex-heavy buildout to sustained earnings traction.