SocGen Recommends 10 Cybersecurity Stocks, Betting on AI Risks to Drive New Demand
nashnova research
SocGen chief U.S. equity strategist Manish Kabra is urging clients to buy cybersecurity, citing 16% compound EPS growth since 2020, valuations still near historical averages, and AI-safety spending extending from compute into governance — a sign the sector may be AI's underpriced second act.
Why is SocGen bullish on cybersecurity now?
Since 2020, cybersecurity-sector EPS has compounded at roughly 16%, up from 10% in the prior decade. This means → the earnings acceleration is not a blip but a five-year structural shift.
Despite the recent rally, the basket trades at a forward P/E of about 25×, still below its 30.2× average since 2018. In plain terms = after a strong run, the sector is not expensive by its own history.
Kabra argues that governments and industry are converging on AI-safety priorities, pushing spending from compute infrastructure toward AI-system security and governance infrastructure.
How strong has recent performance been?
The Global X Cybersecurity ETF (ticker: BUG) rose 10% in a single week.
Palo Alto Networks (PANW) and CrowdStrike (CRWD) gained 10% and 15% over the same period; year-to-date returns now stand at 101.6% and 107.3% respectively.
This reflects the market already pricing in the "AI-safety spending" thesis — short-term momentum is significant.
How have the 10 individual stocks performed?
Four names are up more than 100% year-to-date: Okta (+120.2%), Fortinet (+116%), CrowdStrike (+107.3%), Palo Alto Networks (+101.6%).
Mid-range gains: Cloudflare (+67.2%), Akamai (+22.1%), Gen Digital (+11.4%).
Clear laggards: Zscaler (−13.3%), CyberArk (−9.4%), Check Point (−27.6%). This means → within the same sector thesis, stock-level dispersion exceeds 140 percentage points — picking the right name matters more than picking the right theme.
How does cybersecurity relate to traditional AI trades?
SocGen notes that the cybersecurity basket has shown low — sometimes negative — correlation with semiconductor stocks in 2026.
In plain terms = when chip stocks pull back, cyber names don't necessarily follow, and vice versa — the two move on different rhythms.
This means → for investors already heavy in the AI-compute chain, cybersecurity offers a diversification option rather than doubling down on the same bet.
What is the biggest uncertainty in this thesis?
Intra-sector dispersion is stark: top performer Okta (+120.2%) and worst performer Check Point (−27.6%) sit worlds apart.
This reflects a core question — AI-safety spending is still largely at the expectation stage; whether it converts into actual revenue growth company by company is the make-or-break for the thesis.
Put simply = the direction may be right, but the market has not yet agreed on who will capture the dollars.
市场有风险，内容仅供研究参考，不构成投资建议。