The selloff is overdone. There was an inconsistency last week: the market penalized hyperscalers for AI capex while also believing AI adoption is so quick it would destroy entire SaaS sectors. Only one story can be right, and AI does matter, creating winners and losers.

The risk is concentrated in horizontal SaaS firms catering to SMEs, not in vertical SaaS (banking, gov, pharma) or cybersecurity. Big, regulated clients won't switch to unproven tools due to compliance and security issues.

Earnings risk exists in some SaaS sub-sectors, but not everywhere. 'System of record' for big, conservative companies is sticky; they won't touch existing systems just for a cheaper price if it brings compliance risk.

The US is in the early-to-middle stages of the single biggest industrial investment cycle since WWII, driven by one sector. History rhymes; winners and losers weren't clear for years with e-commerce.

For 2026, not a single hyperscaler is pulling back on spending. They will spend 'hand over fist.' It's very hard for the US economy to dip significantly no matter what Bitcoin does.

It's absolutely possible. Parts of SaaS are being oversold. Hyperscalers are fantastic profit engines; the market questions their spending, not their earnings power. To sell Microsoft, you need a strong view on earnings collapsing, not unjustified AI spend.