When Wall Street begins comparing software stocks to the newspaper industry, it signals that market fears over AI disruption have reached an extreme level. According to trading desk analysis, Goldman Sachs analyst Ben Snider and his team, in a recently published report, made a rare comparison by placing the current software industry alongside the newspaper sector disrupted by the internet in the early 2000s and the tobacco industry hit by heavy regulation in the late 1990s. This analogy alone illustrates how Wall Street is pricing in the risk of "AI disrupting software business models."

Goldman Sachs argues that the current valuation decline does not reflect short-term profit fluctuations but rather fundamental doubts about the software industry's long-term growth and profit margins. The firm warns that when a sector is perceived to face disruptive risks, stock prices bottom out only when earnings expectations stabilize, not when valuations appear cheap enough.

From "AI dividends" to "AI threats": Software stocks face collective reassessment. Goldman Sachs points out that over the past week, software stocks have become the epicenter of the AI disruption narrative. The software sector plunged 15% in a single week, accumulating a 29% decline from its September 2025 peak. Goldman’s "AI Risk Exposure Basket" has fallen 12% year-to-date. Key catalysts for the shift in market sentiment include Anthropic’s release of Claude collaboration plugins and Google’s launch of the Genie 3 model. Investors now view these developments not merely as productivity enhancers but as direct threats to software companies' pricing power, competitive moats, and even their existential value.

Goldman Sachs explicitly states in the report that the market is no longer debating minor earnings revisions but is questioning whether the software industry is headed for a long-term decline similar to that of newspapers.

Valuations appear "rational," but the market is already betting on a growth collapse. On the surface, software stock valuations have declined significantly:

The sector’s forward price-to-earnings ratio has dropped from around 35x at the end of 2025 to about 20x today, hitting its lowest level since 2014. Valuation premiums relative to the S&P 500 have also fallen to their lowest in over a decade.

However, Goldman emphasizes that the problem lies not in valuations but in the breakdown of the assumptions behind them. The report shows that the software industry’s profit margins and consensus revenue growth expectations remain at their highest levels in at least 20 years, significantly above the S&P 500 average. This implies that the sharp valuation decline reflects expectations of a substantial downward revision in future growth and profitability.

Goldman’s comparative analysis reveals:

In September 2025, when software stocks traded at a P/E of 36x, they were priced for mid-term revenue growth expectations of 15%–20%. Today’s P/E of around 20x corresponds to growth assumptions of just 5%–10%.

In other words, the market is already pricing in a "growth cliff."

The warning from the "newspaper moment": Earnings stability, not valuation, marks the bottom. The most striking part of the report is Goldman’s reference to historical case studies. The firm notes that newspaper stocks fell an average of 95% between 2002 and 2009, with the real bottom occurring only when consensus earnings expectations stopped declining—not when macroeconomic conditions improved or valuations became cheap. A similar pattern was observed in the tobacco industry in the late 1990s: despite sharp valuation compression, stock prices remained under pressure until regulatory uncertainty was resolved with the Master Settlement Agreement.

Based on these examples, Goldman’s conclusion is sobering, if not pessimistic: even if short-term earnings show resilience, it may not be enough to dismiss the long-term downside risks posed by AI.

Capital is voting with its feet: fleeing "AI risk" and embracing the "real economy." Amid rising AI uncertainty, market preference is shifting away from "AI risk" and toward the "real economy." Goldman data shows that hedge funds have significantly reduced exposure to the software sector, although they remain net long overall. Meanwhile, large mutual funds began systematically underweighting software stocks as early as mid-last year. At the same time, capital is flowing into sectors perceived as less vulnerable to AI disruption, including industrials, energy, chemicals, transportation, and banking. Goldman notes that its tracked value factor and industrial cycle-related portfolios have significantly outperformed recently.

Despite the cautious tone, Goldman has not turned entirely bearish. Its analysts believe certain sub-sectors still offer defensive qualities:

Vertical software, deeply embedded in industry workflows with high customer switching costs, is less susceptible to direct AI replacement. Information services and business services firms with proprietary data and clear industry barriers may have their AI risks overstated by the market. Some companies highly correlated with software but with non-pure software business models show signs of being oversold.

However, the precondition remains clear: only when earnings expectations truly stabilize can stock prices find a bottom.

If the core narrative for software stocks over the past two years was "AI will amplify growth," Goldman’s report marks a turning point—the market is now seriously debating whether AI will erode the very business value of software. The real question is not whether software stocks can rebound, but which software companies can prove they will not become the next newspaper industry.