Theme

Insurers withdraw from what the technology now does

Explore this theme’s evidence over time →

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to April 27, 2026 (theinnermostloop.substack.com)

When carriers exclude AI-related damages from corporate policies, the losses stop being priced and start being absorbed. Uninsurability is the market’s way of saying the tail is unknown.

Development timeline

Newest reporting dates first. Reporting dates may differ from event dates.

2026-09-17: Bessent rejects an AI liability waiver

2026-08-31: A central banker warns the G20 that autonomous models could reprice cyber risk

2026-08-31: Swarms push kernel CVEs from 500 toward 2,000 per release

2026-08-27: A fund falls 67% on cheap Chinese models as regulators subpoena its lenders

2026-08-13: One hallucinated recipe kills twenty-five acres of crops

2026-08-12: Half a trillion is marshalled as chips become an investable asset class

2026-08-06: A hyperscaler sells $25 billion of bonds into $115 billion of demand

2026-08-05: A fund posts its best month in years buying a rival's wreckage

2026-08-04: An idled copper mine restarts in four months on autonomous software

2026-08-02: A company caps vulnerability reports after AI submissions buckle its reviewers

2026-08-01: Jailbreak costs across frontier models vary a hundredfold

2026-07-31: A browser fixes more security bugs in a month than in 23 prior releases

2026-07-30: An AI-focused fund is forced to unwind its entire public book

2026-07-29: AI begins erasing call centers with half the profession exposed by 2030

2026-07-23: Kernel maintainers face 432 machine-found CVEs in a single weekend

2026-07-20: A hacker wipes a country's land registry and its backups

2026-07-18: Top open models trail the cyber frontier by just four to seven months

2026-07-16: A lab adapts watermarking to biology and points its models at outbreaks

2026-07-15: A security firm turns prompt injection into a defense that derails attackers

2026-07-03: Researchers document the first end-to-end agentic ransomware

2026-06-25: A lab accuses a rival of harvesting 28.8 million exchanges

2026-06-24: A cyber model shifts the mission from finding flaws to patching them

2026-06-21: A former NSA chief says a model cracked nearly every classified system

2026-06-19: Most datacenter capacity sits in markets exposed to climate hazard

2026-06-17: A national agency moves to require quantum-safe cryptography

2026-06-17: A coalition patches 2,000 flaws before disclosure

2026-06-12: A model family is judged seven months ahead of trend on exploit development

2026-06-12: A court rules AI summaries are the platform's own speech

2026-06-07: Critical CVE disclosures spike as a model preview lands

2026-06-03: A defensive model is extended to 150 critical-infrastructure organizations

2026-06-02: A European cyber agency is given access to a model officials fear

2026-05-29: An SSD side-channel fingerprints open apps by disk timing

2026-05-29: IBM and Red Hat pledge $5B to secure the open-source supply chain

2026-05-28: AI bug-finders push kernel CVE issuance to thirteen a day

2026-05-26: Fabricated references in biomedical literature grow twelvefold

2026-05-25: A central bank summons banks over a model's systemic threat

2026-05-24: A portfolio sweep surfaces 10,000+ high and critical vulnerabilities

2026-04-27: Major insurers win approval to exclude AI damages from corporate policies

Knowledge graphopen in browser →

Cite or correct this record

Citation formatCitation text

Insurers withdraw from what the technology now does
Innermost Loop KB — secondary record
Reporting date: unknown
KB permalink: https://nicholsn.github.io/innermost-loop-kb/themes/risk-becomes-uninsurable
Attributed sources:
- Welcome to April 27, 2026 — https://theinnermostloop.substack.com/p/welcome-to-april-27-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.