Anthropic · Friday, October 9, 2026
Anthropic starts a program to help protect power grids, water systems, and open-source code
The File desk · Oct 9, 2026, 10:44 AM UTC
Status
Confirmed. Anthropic's news page, dated Oct. 8, 2026. The description of the program is the company's.
- Confirmed
Anthropic launched the Anthropic Cyber Mission, with a Critical Infrastructure Defense Program and a free scanner for open-source projects.
- Confirmed
The same day Anthropic published a
update that takes effect Nov. 12.Legal instrument
Anthropic 2026 Usage Policy
Anthropic's rules for people using Claude, updated Oct. 8, 2026, and effective Nov. 12. It is a company policy, not a law. The summary post says the rewrite covers deceptive campaigns, weapons, surveillance, and machines Claude might control.
Anthropic launched what it calls the Anthropic Cyber Mission on Thursday, a long-term effort to help defenders secure critical systems. It begins with two pieces. A Critical Infrastructure Defense Program will give its top Claude models, on-site engineers, and threat research to the security firms and equipment makers that protect power, water, and transportation networks. An OSS Scanner will scan opted-in open-source software projects for security holes for free and suggest fixes. Anthropic says frontier artificial intelligence makes it easier to find weaknesses, while fixing them is still hard.
Anthropic's page, dated Oct. 8, 2026, says: “Today we’re launching the Anthropic Cyber Mission, a long-term commitment to securing the systems everyone depends on.” Founding partners named on the page: Accenture, Booz Allen, CrowdStrike, Deloitte, Dragos, Hitachi, Insane Cyber, Nozomi Networks, Palo Alto Networks, PwC, and Rockwell Automation. “Operational technology” means the controllers and industrial networks that run plants and grids. Anthropic says this equipment often cannot be taken offline to patch.
On the scanner, Anthropic says: “Today, we’re launching OSS Scanner, which offers open-source projects regular security scans from our strongest models, for free.” Anthropic says scan reports will include a proof-of-concept, an explanation, and a suggested fix. No dollar figure for the mission is on the page. Earlier this week Anthropic folded a project called Glasswing into an expanded Cyber Verification Program. This site ran that program on Wednesday. It is context, not a new announcement.
The same day Anthropic published a Legal instrument Anthropic's rules for people using Claude, updated Oct. 8, 2026, and effective Nov. 12. It is a company policy, not a law. The summary post says the rewrite covers deceptive campaigns, weapons, surveillance, and machines Claude might control.Anthropic 2026 Usage Policy
What is still unknown or disputed
- The cost and size of the Cyber Mission.
- Which open-source projects have opted in.
- The full Usage Policy text. Only the summary post was opened.
Primary sources
Every claim in this story is drawn from the documents below. If a fetch failed, that is recorded on the card.
Source 1
Introducing the Anthropic Cyber Mission
Anthropic · October 8, 2026
Today we’re launching the Anthropic Cyber Mission, a long-term commitment to securing the systems everyone depends on.
https://www.anthropic.com/news/anthropic-cyber-mission
Source 2
Anthropic on X
Anthropic · October 8, 2026
We’re launching the Anthropic Cyber Mission, a new effort to secure critical infrastructure and open-source software.
https://x.com/AnthropicAI/status/2108302539498414208
Source 3
2026 Usage Policy update
Anthropic · October 8, 2026
The updated policy takes effect on November 12.
https://www.anthropic.com/news/2026-usage-policy-update