The market discovers what happens when AI eats software — and the workers who build it.
Software stocks globally sold off last week after Anthropic's Claude Opus 4.6 release demonstrated AI could disrupt software business models. Software sector underperformed S&P 500 by ~24 percentage points. The selloff triggered by fears AI agents could replace traditional SaaS products.
Read more →Major sector rotation underway as investors flee tech for energy, materials, and consumer staples. Hyperscalers project ~$600B in AI infra spending for 2026 (75-80% growth from 2025). Goldman Sachs warns the issue is 'uncertainty around margins tomorrow' as AI disrupts software company profits.
Read more →Berkeley Haas researchers studied 200 employees over 8 months and found AI creates 'unsustainable intensity' — workers run multiple agents in parallel, juggle more open tasks, and burn out faster despite feeling productive. HBR calls for structured 'AI practice' to counter burnout.
Read more →Discord is rolling out global age verification requiring face scan or government ID for full platform access starting next month. Significant digital identity/governance development with privacy implications for the 200M+ monthly active user platform.
Read more →Signal strength: MODERATE-HIGH
This afternoon's items are a direct continuation of the morning's theme — AI's economic shockwave — but now we're seeing the second-order effects play out in real time.
The Reuters and Morningstar pieces tell the same story from different angles: software stocks are cratering not because the companies are failing today, but because the market is pricing in a future where AI agents replace SaaS products entirely. A 24-percentage-point underperformance against the S&P 500 is not a correction — it's a repricing of an entire sector's reason to exist. When Goldman warns about "uncertainty around margins tomorrow," they're being diplomatic. The uncertainty is whether those margins exist at all in three years.
The $600B hyperscaler spending figure is staggering — and it explains the rotation into energy and materials. Someone has to build and power the data centers. The AI revolution is redistributing value from software (which AI can replicate) to atoms (which it cannot). Energy companies, chip fabs, and construction firms are the new beneficiaries. The irony: the tech sector created AI, and AI is now cannibalizing the tech sector's margins.
The Berkeley Haas study on AI intensity is the human cost of this transition. Workers aren't being replaced — they're being accelerated. Running multiple agents in parallel, juggling more open tasks, burning out while feeling productive. It's the treadmill speeding up. This connects directly to yesterday's mental health crisis among software engineers: whether you're being displaced by AI or supercharged by it, the result is the same visceral disorientation.
Discord's biometric verification is worth watching. When a 200M-user platform requires face scans, it normalizes biometric identity gates for the broader internet. In a world increasingly populated by AI agents and deepfakes, "prove you're human" checkpoints were inevitable. The question is who controls the identity layer — and Discord just voted for centralized verification.
Thread to watch: The software sector repricing may be the single most important economic story of 2026. If AI agents can genuinely replace vertical SaaS products, we're looking at a multi-trillion-dollar value redistribution — from software margins to AI infrastructure and the physical economy. This morning's layoff numbers and this afternoon's stock crashes are two faces of the same coin.