AI has killed software and tech stocks. Valuations are at multi-decade lows. Stock prices are in the trash. This is a once in a lifetime opportunity to buy fear. Follow as I uncover hated tech stocks trading at basement level valuations.
Contains line charts (infographics) of median EV/Revenue and EV/EBITDA for software companies from H1 2015–H1 2025 showing a peak in 2021 followed by a sharp correction to roughly ~3x EV/Revenue and ~17–18x EV/EBITDA by H1 2025 — directly illustrating compressed software/tech valuations and supporting the claim that tech/software valuations have fallen dramatically (i.e., a once-in-a-cycle buying opportunity).
Source: Aventis Advisors
Research Brief
What our analysis found
In early February 2026, Anthropic's release of new Claude "Cowork" agentic plugins and coding capabilities triggered one of the sharpest selloffs in software stocks in recent memory. According to Axios, over $400 billion in market capitalization was erased from the software sector in a single week, with the sector plunging roughly 25% in that span. Individual casualties were severe: Oracle shed $61 billion in market cap, Palantir lost $36.1 billion, and AppLovin dropped $22.6 billion. Some commentators aggregated multi-day losses as high as $1 trillion, though estimates vary widely depending on the time window and methodology used.
The carnage extended across indices and geographies. The S&P North American Technology Software Index fell more than 20% year-to-date through February 6, 2026, according to S&P Global Market Intelligence, while iShares Expanded Tech-Software ETFs hit their lowest closes since 2023–2024 ranges. Single-day moves were dramatic: AppLovin tumbled 19.7%, Cisco fell 12.3%, and IBM suffered a roughly 13% plunge on February 24 after Anthropic showcased COBOL modernization capabilities — IBM's worst single-session decline in decades. The episode was widely dubbed the "SaaSpocalypse" across financial media.
Valuation compression accompanied the price declines. iCapital reported the S&P software industry forward price-to-earnings multiple contracting to around 23.7x, the lowest level since 2022 and in the bottom percentile of its ten-year range. Some analysts noted the software index briefly traded below 20x forward earnings. While the tweet's claim that valuations sit at "multi-decade lows" stretches the data — multiples are at multi-year lows rather than truly generational troughs — the scale and speed of the repricing are well-documented and historically unusual for the sector.
Fact Check
Evidence from both sides
Supporting Evidence
Axios documented $400 billion erased in one week
Axios reported on February 7, 2026 that Anthropic's Claude Cowork release triggered a sector-wide revaluation, wiping over $400 billion from software stocks in a single week, with the sector declining approximately 25%. The report itemized specific company losses including Oracle at $61 billion, Palantir at $36.1 billion, and AppLovin at $22.6 billion.
Global selloff confirmed by Reuters and multiple outlets
Reuters and other press reported that global software stocks — spanning both U.S. and European names — sold off sharply after Anthropic's agentic AI plugins were publicized in early February, confirming the breadth of the market reaction across geographies and sub-sectors.
S&P Global measured a greater than 20% index decline
S&P Global Market Intelligence documented that the S&P North American Technology Software Index was down more than 20% year-to-date through February 6, 2026, and explicitly linked the selling pressure to Anthropic and OpenAI coding tool announcements, noting fundamental risk repricing across the sector.
Successive Anthropic announcements caused repeated waves of selling
The selloff was not a single event. Bloomberg and Axios reported that cybersecurity and software ETFs hit multi-month lows again on February 20–23 when Anthropic announced Claude Code Security capabilities, demonstrating that each new product drop triggered an additional leg down in affected sub-sectors.
Forward valuation multiples compressed to multi-year lows
iCapital reported the S&P software industry forward multiple fell to approximately 23.7x, the lowest since 2022 and in the bottom percentile of its ten-year range, supporting the claim that valuations experienced significant compression driven by AI disruption fears.
Contradicting Evidence
Analysts called the selloff an overreaction
Multiple sell-side and buy-side analysts quoted by S&P Global and Axios argued the market was extrapolating worst-case AI disruption scenarios into near-term earnings damage far beyond what fundamentals warranted. Dan Ives of Wedbush said the market was "baking in a doomsday scenario," and Raymond James called the cybersecurity reaction "excessive," suggesting the fear-driven repricing may not reflect durable structural change.
"Multi-decade lows" overstates the valuation picture
While forward multiples did compress meaningfully, the data shows software valuations at their lowest since 2022 — a multi-year low, not a multi-decade low. The tweet's framing of "multi-decade lows" and "basement level valuations" is more dramatic than what the available data from iCapital and S&P Global supports, as the sector traded at far lower absolute and relative multiples during periods like the early 2000s dot-com bust.
Nvidia CEO pushed back on the disruption narrative
Jensen Huang publicly dismissed the idea that AI would replace the software industry, calling the view "the most illogical thing in the world" at a Cisco AI event on February 4, 2026. His argument — that AI will expand and rely upon existing software infrastructure rather than instantly destroy incumbents — challenges the premise that traditional software companies face existential threats warranting such steep markdowns.
The tweet frames a complex event as a simple buying opportunity
The viral post presents the selloff as a straightforward contrarian buying signal, but S&P Global's analysis noted that the repricing also exposes deeper structural concerns about software growth rates, employee compensation models, and long-term competitive positioning against AI-native tools — suggesting the discounted valuations may partly reflect legitimate fundamental risks rather than pure irrational fear.
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.