News - Two days in February. Same AI technology. Palantir posts +70% revenue growth. The software sector watches $285 billion evaporate.
• Same AI, opposite outcomes: Palantir surges, software stocks crash. • The difference is context. • AI rewards vendors embedded in unique, defensible data and workflows. • It commoditises generic SaaS fast.
Connections
Two days in February. Same AI technology. Palantir posts +70% revenue growth. The software sector watches $285 billion evaporate.
I find this contrast from February 2nd-4th striking. It connects directly to what I've been writing about context.
Palantir announced Q4 2025 revenue of $1.41 billion with net income soaring from $79 million to $608 million [1]. Two days later, Anthropic released Claude Cowork plugins targeting legal and professional workflows, and the software sector cratered. Thomson Reuters dropped 18%. RELX fell 14%. LegalZoom lost almost 20% [2]. Goldman Sachs' software basket had its worst single day since April's tariff selloff.
Both stories involve AI. The market reactions were opposite.
Palantir solves problems foundational models cannot touch: classified intelligence networks, defence targeting systems, proprietary data pipelines that Google and Anthropic legally and technically cannot replicate. Their moat is context—the specific, sensitive kind that no general-purpose API can automate.
The threatened SaaS companies solve problems AI platforms can now replicate at marginal cost. Organising case law, drafting standard contracts, automating professional workflows. Claude can read the same documents and produce similar outputs for a fraction of the price.
For CFOs and FinOps practitioners, this changes the approval question. Before signing off on AI or SaaS spend, you need to ask whether this solves complex, context-specific problems unique to your organisation, or generic problems AI will commoditise.
For FinOps teams, this changes what a SaaS portfolio audit means. The old audit asked about utilisation. The new audit asks about defensibility against AI commoditisation and what happens to a vendor's margin when foundational models automate its core function.
Your SaaS portfolio likely contains both categories: tools that embed into your unique workflows and data, and tools that solve generic problems any AI platform could replicate tomorrow. Knowing which is which has become strategic work.
Context is everything. And when it's absent, markets notice fast.
Sources:
[1] Palantir Q4 2025 earnings report, February 2026
[2] Times of India, Anthropic AI Tool Impact, February 2026
I explore context-driven FinOps frameworks at frankcontrepois.com
Note: Initial draft assisted by AI with my fine-tuning.