Software stocks got crushed. Did they have it coming?

Episode Summary

Executive Summary: The episode examines whether the sharp selloff in software stocks after AI-related panic was a rational repricing or an overreaction. The hosts argue that mission-critical enterprise software is sticky and unlikely to be quickly displaced, but also warn that AI enthusiasm is concentrated in a tiny set of stocks and could be vulnerable to a bubble unwind.

Main Topics: AI panic and the February software-stock selloff (Priority: 5/5): The hosts revisit the market shock sparked by fears that AI would disrupt software companies and broader employment, leading investors to punish enterprise software names heavily. Why enterprise software may be more durable than the market implies (Priority: 5/5): They argue that core business software is deeply embedded in corporate operations, compliance, and data workflows, making it difficult and costly to replace quickly. The SAS pocalypse: valuation compression in software (Priority: 4/5): Major software-as-a-service companies are described as trading far below recent highs and at valuations more akin to cyclical or distressed businesses, despite ongoing growth. AI’s uncertain impact on software value capture (Priority: 4/5): A possible split is discussed between AI commoditizing the user-interface and application layer while leaving the underlying data layer intact, versus AI hype itself proving excessive. Concentration risk in AI-driven equity markets (Priority: 5/5): The Bank of England’s warning highlights how a small number of AI-related stocks now account for a huge share of market capitalization, creating systemic vulnerability if sentiment reverses. Market overshoot, central-bank warnings, and practical investor limits (Priority: 3/5): The hosts note that markets routinely overshoot and undershoot, but central-bank scenario analysis still serves as a cautionary signal even if it is not directly actionable for investors. Long/short segment: dissenting analyst coverage and egg-price integrity (Priority: 2/5): Rob praises a lone sell-side analyst shorting SpaceX, while Katie calls out alleged egg-price collusion as a benchmark/data integrity scandal that distorted inflation readings.

Key Arguments: The February selloff in software stocks was driven by a market search for reasons to panic about AI, not by clear evidence that enterprise software demand had already collapsed. Mission-critical software is hard to rip out because it supports payments, data workflows, regulation, and compliance; switching costs are high and incumbents like Oracle have survived prior obsolescence scares. Software stocks may have been repriced too aggressively, with some names falling to valuations typical of banks or weak cyclical businesses despite continuing revenue growth. AI may ultimately erode the value of the application/user-interface layer more than the underlying data infrastructure layer that enterprise software providers control. The broader AI trade is highly concentrated, with a small number of chipmakers, hyperscalers, and AI-linked names driving index performance and creating fragile market leadership. Central banks are warning because they cannot know whether AI will produce a bubble, a crash, or a structural transformation, but they want to flag the risk before it is obvious. Investors need diversification because even if a 45% market drop is extreme, drawdowns of around a third are not unusual over long horizons.

Data Points: Software stocks performance: Down about 20% year over year - Katie says software stocks remain depressed even as the broader market has risen. S&P 500 Software and Services Index: Down 21% from a year ago - Used to illustrate the scale of the software-sector drawdown. Peak-to-current decline for some software names: Trading at 40% or even a third of recent highs - Rob describes the magnitude of valuation compression in large enterprise software firms. Global equity market share of the S&P 500: About half - From the Bank of England’s concentration-risk discussion. AI companies’ share within the S&P 500: About half, up from about a quarter in 2022 - Bank of England warning about increasing concentration in AI-related stocks. Hypothetical market shock: 45% drop in the US stock market over six quarters - A scenario outlined by the Bank of England to test financial stability under an AI-related downturn. Typical market drawdown: A third, a couple times per decade - Rob notes that large drawdowns occur periodically and should be planned for regardless of AI.

Pivotal Quotes: "what we want to figure out is whether that market shakeout was overdone or a warning of things to come" — Katie Martin: Framing the episode’s central question about AI-driven software stock declines. "You don't just like switch that off and plug your business into Claude. No. That's just ridiculous." — Katie Martin: Explaining why mission-critical enterprise software is hard to replace quickly with AI tools. "the entire global stock market is balancing on this tiny, tiny pinhead that is like a tiny number of AI stocks" — Katie Martin: Summarizing the Bank of England’s concentration-risk warning.

Implications: Listeners should view AI as both a real competitive threat and a source of market overreach. Enterprise software may be stickier than feared, but the AI trade’s concentration makes the market vulnerable if sentiment shifts.