Is it time to buy growth stocks?

Maybe I should rephrase in light of new records for large-cap US equity benchmarks which tend to be dominated by growth shares: Was it time to buy growth stocks?

First we have to define “growth stocks.” Goldman has a “Rule of 10” which defines a secular grower as a non-financial, non-real estate, non-utility S&P company which i) grew sales by 10% or more over the past two years, ii) is on pace to do so in the current year and iii) will likely replicate the feat in each of the next two years, according to company analysts.

Plainly, that list changes dynamically, but Goldman runs an “evolving screen” for such companies. They’ve underperformed since October to a degree rarely seen on a 15-year lookback.

The figures below give you a sense of absolute and relative six-month performance. On the left, you can clearly see the impact of the AI disruption scare in late-January/early-February, followed by a partial recovery and then another leg lower as the bombs started falling on Iran.

The chart on the right gives you a sense of how 2026’s “double whammy” (“SaaSpocalypse” and higher bond yields as oil prices surged) erased two years of relative outperformance.

If you’re curious, there were nearly three-dozen S&P stocks which met Goldman’s secular growth criteria headed into this week, the most ever. Almost a third of them were software companies and “a number of others” had significant “exposure to AI investment and uncertainty,” as the bank’s Ben Snider put it.

The question’s whether — checks notes — almost 30ppt of underperformance since October versus the equal-weighted S&P counts as overdone. Headed into this week’s summit push on Wall Street, secular growth stocks on Goldman’s definition were still down 20% versus just 2% for the broader, cap-weighted index.

Moreover, the de-rating in the highest-growth shares is quite dramatic: 30% in fact (on the left). And Goldman’s basket trades at “just” 27x (on the right).

No, 27x isn’t “cheap” on any conventional definition, but it’s bargain in the post-pandemic context. The premium to the median stock is less than 50%, among the lowest of the post-GFC era.

This isn’t just a software phenomenon, by the way. So, the analysis isn’t irrelevant even if you’re all-in on the software apocalypse narrative (as a quick aside, my own recent “conversations” with Claude suggest that narrative’s overdone, or at least if basic math errors on “his” part are any indication).

“Even excluding Software, many secular growth stocks have recently underperformed and trade at discounted valuation multiples relative to the past decade,” Snider went on, noting that the median non-Software secular grower trades at 29x. The premium to the median S&P stock, 53%, “is close to the bottom of the range during the past 10 years.”