Goldman Sachs says the AI capex cycle is in its early innings and expects leadership to rotate away from the mega-cap winners of the first rally toward new beneficiaries, even as its Delta One desk warns that falling AI token prices could pressure equity valuations.
Goldman Sachs is telling clients the artificial-intelligence spending cycle is barely past its opening phase, even as the stocks that powered the first leg of the rally face a changing of the guard. The bank's strategists argue the buildout of AI infrastructure has years of capital deployment ahead, but the names that delivered the biggest gains so far may not be the ones that lead the next stretch of the trade.
"The AI trade is in its early innings, and the stocks that got you here will not take you there," the Goldman Sachs team wrote in a note published this week, according to the firm's research. The call frames the current moment as a transition point in a multiyear investment cycle rather than the start of a retreat from AI exposure.
The thesis rests on a distinction between the first wave of AI winners and the beneficiaries of what comes next. Goldman's strategists expect leadership to rotate toward parts of the value chain that stand to gain as the cycle matures, a shift that could redirect flows away from the mega-cap names that dominated the initial surge. The bank did not name specific replacement beneficiaries in the note, leaving investors to parse which segments of the AI economy capture the next round of spending.
A separate warning from Goldman's Delta One desk adds a cautionary layer to the bullish framing. Falling prices for AI tokens threaten to undermine equity valuations unless computing consumption surges enough to offset the decline, the desk cautioned in an institutional note. Per-token pricing may not hold as AI inference shifts toward cheaper, localized hardware, making it harder for providers to sustain monetization under competitive price pressure. The industry could face a glut in infrastructure and stretches of compute oversupply if output costs fall faster than consumption grows, a scenario the desk said is not automatically bullish for investors.
The two messages together sketch a cycle that is young but not without risk. Goldman's core view is that capital spending on AI remains in its early phase, which should keep the broad trade alive. Yet the path forward is unlikely to mirror the past, with the bank pointing to rotation rather than repetition as the defining feature of the next leg.
For investors, the takeaway is a question of positioning rather than conviction. The firms and hyperscalers that poured capital into the first wave of AI buildout — the cohort that drove much of the market's advance — may cede leadership to names better placed for the next phase of deployment. Goldman's strategists have not yet published the full list of expected beneficiaries, meaning the specific calls will matter more than the broad thesis in the weeks ahead.
This article is for informational purposes only and does not constitute investment advice.
