Research
Goldman Sachs measures the AI momentum trade's volatility at a 45-year high outside recessions, and points research desks at three ways around it
Goldman's research desk puts three-month annualised volatility in the momentum factor at 36 percent, the highest reading outside a recession in the factor's 45-year history, and flags consumer-experience names, quality compounders and M&A targets as the least AI-correlated places to look instead.
Goldman Sachs's research desk has put a hard number on something that traders have been describing anecdotally all year: the AI momentum trade has become genuinely unstable, not just directionally uncertain. The bank measures three-month annualised volatility in the momentum factor at 36 percent, which it flags as the highest level outside of a recession in the factor's 45-year history. That is a meaningful distinction from ordinary drawdown risk. It means the AI-momentum cohort is now swinging with a violence normally reserved for genuine macro shocks, recessions and crises, even though the broader economic backdrop has not deteriorated to match.
The practical response Goldman's report offers is a diversification map away from AI-correlated names rather than a call to exit the trade outright, and the three themes it names are deliberately chosen for their low correlation to the momentum factor rather than for being anti-AI. Consumer experience companies, businesses tied to travel, entertainment and leisure spending, are flagged as relatively insulated because that demand for physical experiences has stayed resilient through the AI volatility and does not move on the same triggers. High-quality compounders are the second theme, chosen specifically because their growth profile does not lean on the AI or momentum factors that are currently whipsawing. The third is M&A targets, backed by a concrete data point: announced US M&A activity has reached roughly $1.2 trillion year to date, up 32 percent from a year earlier, with deal count itself up 12 percent, evidence that capital is actively rotating into a channel that runs on corporate strategy rather than AI-adjacent price momentum.
The volatility measurement itself is the more durable finding here, independent of whether Goldman's specific three themes hold up. A 45-year record outside recession conditions is a statement about market structure, not just about one factor having a rough quarter, and it implies that risk models built on momentum's historical volatility profile are currently understating tail risk in anything sized against that factor. Desks running factor-based risk budgets that have not explicitly re-parameterised for this regime are likely carrying more exposure than their models show.
The caveat is that a volatility spike this sharp inside a single factor is also consistent with a temporary crowding unwind rather than a permanent regime change, and Goldman's own three-theme diversification call is itself a forecast, not a guarantee that consumer, quality or M&A names stay uncorrelated to AI-driven momentum once enough capital rotates into them on the same thesis. The M&A data point in particular deserves scrutiny on its own terms since elevated deal activity can just as easily reflect AI-driven strategic consolidation as an escape from it. Treat the 36 percent volatility read as a real, well-measured warning about current momentum-factor risk, and treat the specific rotation targets as one bank's reasoned bet rather than a settled diversification formula.
Read the original: Seeking Alpha - Goldman Sachs measures the AI momentum trade's volatility at a 45-year high outside recessions, and points research desks at three ways around it. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.