Goldman Sachs: Long-duration growth stocks have clearly de-rated, but AI uncertainty demands selectivity

Overview

This is a Goldman Sachs U.S. portfolio strategy report focused on the de-rating of long-duration growth stocks. The report notes that although the S&P 500 has rebounded to near record highs, stocks that meet the long-duration growth screen still lag meaningfully, mainly because of style rotation driven by expectations of an accelerating economy, rising bond yields, and AI disruption risk. The authors argue that the current valuation discount, combined with forecasts for milder growth and lower yields, supports long-duration growth stocks, but AI-related uncertainty will continue to pressure some industries.

Core views

The core view is that long-duration growth stock valuations have compressed significantly, and valuations and PEGs for some non-software long-duration growth stocks are near historical lows, making them relatively attractive. However, AI disruption risk will not disappear quickly, especially for software and other industries whose business models may be displaced by AI, creating sustained discount-rate pressure. Therefore, investors should not simply buy back all long-duration growth stocks broadly, but instead focus on areas with clearer fundamental growth visibility and more explicit benefits from AI investment, especially power infrastructure and resilience-related stocks.

Analysis framework

The report combines stock screening, valuation percentile analysis, relative performance comparison, macro scenario analysis, and thematic basket analysis. First, it uses the Rule of 10 to screen S&P 500 companies that have sustained and are expected to sustain at least 10% sales growth. Second, it compares the performance of long-duration growth stocks with equal-weighted S&P 500, cyclical, defensive, long-duration, and short-duration baskets. Third, it explains style performance differences using U.S. GDP growth, the 10-year Treasury yield, and AI adoption progress. Fourth, it updates the Long Duration GSTHLDUR and Short Duration GSTHSDUR baskets.

Methodology notes

• Stock screeningRule of 10

  Long-duration growth stock screening rule

  Screens S&P 500 companies excluding financials, real estate, and utilities whose sales growth has been at least 10% in each of the past two years and whose consensus expected sales growth for the current year and the next two years is also at least 10%.

• Valuation analysisPEG ratio

  Growth-adjusted valuation

  The report defines PEG as forward P/E relative to consensus sales growth over the next three years, and uses it to judge the valuation position of long-duration growth stocks after adjusting for growth expectations.

• Portfolio strategyLong Duration GSTHLDUR / Short Duration GSTHSDUR

  Equity duration baskets

  Goldman estimates the cash-flow duration distribution of Russell 1000 stocks, constructs long-duration and short-duration baskets, and keeps them industry-neutral relative to the Russell 1000.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Rule of 10 long-duration growth stocks

  Core research object

  Strengths

  Valuations have compressed sharply, sales growth expectations still remain well above the S&P 500 median company, and slower macro growth plus lower rates may provide support.

  Weaknesses

  They have lagged significantly over the past six months, some industries face AI disruption risk, and discount rates may stay higher.

  Comparison

  Underperformed the equal-weighted S&P 500 by about 27 percentage points over the past six months; forward P/E fell from 36x to 27x.

  Risks

  AI replacement of business models, renewed rise in yields, downward revisions to growth expectations.

• Software long-duration growth stocks

  The industry most concentrated with AI disruption risk

  Strengths

  They still have high-growth characteristics and account for a high share of the long-duration growth screen.

  Weaknesses

  AI tools and application launches have increased market concerns that software business models may be displaced.

  Comparison

  Software is one of the areas where de-rating is most pronounced among long-duration growth stocks.

  Risks

  AI products reduce customer demand for traditional software, leading to lower long-term growth estimates.

• Non-software Rule of 10 stocks

  The long-duration growth subset the report sees as more valuation-worthy

  Strengths

  Median P/E of about 29x, close to historical averages and the low end of the past decade; PEG around 1.8x, close to the trough levels seen in 2020 and 2022.

  Weaknesses

  Median performance year to date is only flat, still weaker than the S&P 500 median stock by about 3%.

  Comparison

  Valuation premium to the S&P 500 median stock is about 53%, near decade lows.

  Risks

  Even after excluding software, many companies still face AI risk or high-duration valuation risk.

• Power infrastructure and resilience investment-related stocks

  Preferred long-duration growth theme in the report

  Strengths

  Benefit from power infrastructure demand tied to AI investment, as well as global power infrastructure investment potentially encouraged by Middle East conflict.

  Weaknesses

  Some stocks have already outperformed significantly year to date, so valuation and crowding need to be watched.

  Comparison

  Unlike many other long-duration growth stocks, this group has significantly outperformed year to date.

  Risks

  A slowdown in AI capex, delayed infrastructure investment, policy shifts, and energy price volatility.

• Long Duration GSTHLDUR and Short Duration GSTHSDUR

  The themed baskets updated and rebalanced in the report

  Strengths

  Used to express portfolio views on rate sensitivity and cash-flow duration.

  Weaknesses

  Long-duration stocks are more vulnerable when yields rise sharply.

  Comparison

  From late February to late March, when the 10-year Treasury yield rose to 4.4%, the Long Duration basket underperformed the Short Duration basket by about 1.3 percentage points.

  Risks

  Continued yield increases or a worse inflation path would keep pressuring long-duration assets.

Key data

• S&P 500 positionAbout 2% below the all-time highThe report says ceasefire headlines pushed the index close to its January record high.
• Long-duration growth stock declineMore than 20% below the October 2025 highDespite the broad market rebound, long-duration growth stocks have not recovered.
• Relative performance over the past six monthsAbout 27 percentage points behind the equal-weighted S&P 500One of the worst relative-performance stretches in the past 15 years.
• Valuation compressionMedian stock P/E compressed by about 30%Long-duration growth valuations have clearly de-rated over the past few months.
• Rule of 10 forward P/EFrom 36x to 27xAround the 35th percentile since 2010.
• U.S. GDP forecast2026 Q4/Q4 lowered from 2.5% to 2.0%; about 1.7% in the second halfSlower growth is usually more favorable for stocks with independent growth characteristics.
• 10-year U.S. Treasury yieldRose 47 bp from late February to late March to 4.4%; year-end forecast is about 4.1%Rising yields pressure long-duration growth stocks, while a future decline could improve the backdrop.
• Number of Rule of 10 stocks35 stocksThe current number of qualifying stocks is a record, including 11 in software.
• Software shareAbout 30%Software has become an important component of the Rule of 10 long-duration growth screen.
• Non-software Rule of 10 valuationMedian P/E of about 29x, about 53% premium to the S&P 500 median stock, PEG about 1.8xAll are near the low end of the past decade.
• Power theme stocks54 U.S. power stocksGoldman Sachs equity analysts believe these companies benefit from power infrastructure demand tied to AI investment.

Impact & implications

The investment implication is that long-duration growth stocks have moved from a pure valuation-compression trade into a phase that requires more fundamental validation. If U.S. growth slows but remains positive, and Treasury yields gradually decline, long-duration growth stocks could receive relative support; however, the potential for AI to displace business models means software and some high-growth industries may remain under pressure. On the portfolio side, it is more appropriate to look for companies with valuations that have already come down, growth that can still be delivered, and clearer benefits from AI capital spending or power infrastructure investment, rather than simply betting on a broad rebound in all high-growth stocks.

Risks

• AI disruption risk may persist for several quarters or even years, especially affecting software and other industries whose business models may be displaced.
• If the U.S. inflation path worsens further, the market may continue pricing in a hawkish Fed stance, pushing Treasury yields higher again.
• If sales growth expectations for long-duration growth stocks are revised down, the currently seemingly low P/E and PEG may lose support.
• Middle East conflict, oil prices, and macro uncertainty may affect GDP growth, corporate costs, and risk appetite.
• Theme-basket trading is constrained by market liquidity and stock-borrow availability, so actual tradability may differ from the conditions at the time of publication.

What to watch

• Whether the 10-year U.S. Treasury yield declines toward 4.1% as Goldman Sachs rates strategists expect.
• Whether U.S. real GDP growth in the second half slows to a range close to trend but still positive.
• The real impact of AI adoption on revenue, renewals, pricing power, and long-term growth expectations in software and high-growth industries.
• The sales-growth delivery and earnings revision trends of non-software stocks on the Rule of 10 list.
• Orders, capex, policy support, and valuation crowding in power infrastructure-related stocks.
• Relative performance of the Long Duration GSTHLDUR and Short Duration GSTHSDUR baskets amid rate changes.