Software Sector Rebounds Nearly 40% from "SaaSpocalypse" Bottom
nashnova research
U.S. software stocks have bounced nearly 40% from their April lows, largely erasing the AI-fear selloff; Jefferies sees data platforms and cybersecurity outperforming the application layer, with stock-level divergence ahead.
How big is this bounce?
The iShares Expanded Tech-Software ETF (IGV) surged nearly 40% from its April low — roughly the same magnitude it fell from its Q3 2025 peak. This means → the sector traced a deep V, recovering almost exactly what it lost.
State Street's XSW Software ETF followed a similar path and hit a new all-time high in late August, surpassing its pre-AI-panic level.
In plain terms = early this year the market feared AI would kill traditional software — coining the crash "SaaSpocalypse." That thesis has, for now, been wrong.
Why weren't software companies replaced by AI?
Jefferies analyst Brent Thill argues frontier AI labs are more likely to partner with leading software vendors than build full-stack alternatives.
The moats are threefold: systems of record, proprietary data, and sticky workflows — none of which AI labs can replicate quickly.
This reflects a broader shift: the software-vs-AI narrative is moving from "displacement" to "co-existence." AI needs software as a delivery channel; software needs AI as a feature upgrade.
Which sub-sectors are leading?
Jefferies sees AI monetization advancing faster in databases and operating systems than in consumer-facing apps. This means → data platforms and cybersecurity currently outperform the application layer.
Top picks: data management firm Snowflake, analytics platform Dynatrace, cybersecurity names Palo Alto Networks and Okta.
Snowflake jumped 22% Wednesday after a Q2 beat: adjusted EPS $0.62 vs. the $0.45 LSEG consensus; revenue $1.55 billion vs. $1.48 billion expected. The company also raised its full-year product revenue guidance.
Palo Alto Networks posted a similar Q4 beat: adjusted EPS $1.02 vs. $0.98; revenue $3.41 billion vs. $3.35 billion.
Is the application layer a total write-off?
Monetization across the app layer remains uneven, but outliers exist. Medical-records platform Doximity doubled in after-hours trading this month.
CEO Jeffrey Tangney said on the Q1 call: "Our revenue per search is more than 10 times the cost." In plain terms = this company has already proved AI can generate direct profit at very high margins.
Legacy vendors are also signaling alignment. Salesforce CEO Marc Benioff appeared alongside Anthropic CEO Dario Amodei in late August, stressing complementarity. Amodei said: "We have no intention of disrupting anyone."
What to watch next?
The software recovery thesis is taking shape: infrastructure names are delivering earnings beats, and select app-layer names are breaking out.
But uneven monetization across the application layer means stock-level divergence will persist — not every software company can turn AI into revenue.
This means → the key test in coming earnings seasons is a single question: can infrastructure and data platforms keep converting AI tailwinds into actual results? The answer determines whether this rally is a true reversal or a dead-cat bounce.
市场有风险，内容仅供研究参考，不构成投资建议。