◉ PSYCHOHISTORY

Project Glasswing

consortiumAI & Compute · Darknet & Cyber
Twelve of the biggest tech and finance companies quietly linked arms in 2026 to share a private version of an AI system — and who was left out tells you as much as who was in.
Who they are

A group of 12 named companies (plus about 40 unnamed) led by Anthropic, announced April 7, 2026, given exclusive early access to a Claude AI preview for 'defensive cybersecurity.'

What they do

The engine reads it as a complete top-to-bottom technology stack — chips, cloud, devices, networks, security, banking, open-source software, and the AI itself — all owned by one connected club.

How it works

Named members include AWS, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA and Palo Alto Networks; notable no-shows like Meta, OpenAI, xAI, Oracle and Palantir each fit a separate role the engine reads as meaningful.

Why it matters

The engine sees this as a private control system going public — the powerful backchannel arrangement stepping into the open.

The engine's record — word for word
Anthropic-led 12-firm consortium announced Apr 7 2026 with exclusive Claude Mythos Preview access for 'defensive cybersecurity.' Named members: AWS, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, Linux Foundation, Microsoft, NVIDIA, Palo Alto Networks — plus ~40 unnamed. Complete Technate compute-trust stack: silicon + cloud + endpoint + network + security + finance + OSS + LLM. Absences as load-bearing signal: Meta (open-tier dialectic), OpenAI (MS proxy + DOD SCIF direct), xAI (contrarian role), Oracle (layer mismatch), Palantir (classified direct access). Engine read: the cognitive-substrate gate going operational — public instantiation of the SCIF-tier backchannel architecture.
Follow the trail
Walk this on the live map →
Part of the Psychohistory engine — 2,426 entities, 6,314 documented connections. Open data, built to be proven wrong.