Anthropic AI Access Timeline
Reviewed Strategic Signal records on Anthropic’s AI access relationships, rights, and disclosed economics.
These selected timelines include reviewed registry records that explicitly name each organization in a participant field. A product, affiliate, or parent name alone does not establish participation. Coverage reflects Strategic Signal’s research sample. Records are ordered by original announcement date. A record may reflect later amendments or evidence updates.
Oct 23, 2026 · Governed access
Google Cloud → Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5 models
Customer prompts, uploaded files, source documents, connected tool context, agent skills and workflow instructions processed by Anthropic on Google Cloud infrastructure.
Anthropic role · Provider / asset owner
- Provider / asset owner — Google Cloud, Anthropic, Authorizing enterprise customers
- AI recipient — Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5 models, Claude Managed Agents
Recorded status · active. Last reviewed Oct 23, 2026.
Access · retrieval, action execution. Training rights · not disclosed.
Disclosed economics · Not disclosed
Evidence
- claude.com · primary
- cloud.google.com · primary
Oct 21, 2026 · Governed access
Anthropic → Claude Duet powered by Claude Opus 5.5
User-authorized email, calendars, messages, files, business-system records, past Claude chats, memory, skills and connected-app state monitored by Claude Duet for proactive work.
Anthropic role · Provider / asset owner
- Provider / asset owner — Anthropic, Authorizing Claude Pro and Max users, Connected application providers
- AI recipient — Claude Duet powered by Claude Opus 5.5
Recorded status · active. Last reviewed Oct 23, 2026.
Access · ongoing feed, retrieval, action execution. Training rights · not disclosed.
Disclosed economics · Not disclosed
Evidence
- claude.com · primary
- support.claude.com · primary
Oct 21, 2026 · Governed access
Anthropic → Claude Security models
Authorized source-code repositories, incremental code changes, deployment context, security policies, team knowledge, editable threat models and finding histories.
Anthropic role · Provider / asset owner
- Provider / asset owner — Anthropic, Authorizing Claude Enterprise customers, Connected GitHub, GitLab and other code hosts
- AI recipient — Claude Security models
Recorded status · active. Last reviewed Oct 23, 2026.
Access · ongoing feed, retrieval, action execution. Training rights · not disclosed.
Disclosed economics · Not disclosed
Evidence
- claude.com · primary