Anthropic launches cybersecurity partnership with Nvidia, Cisco and others

Anthropic has launched a cybersecurity partnership with several technology giants, including Amazon, Apple, Broadcom, Cisco, Nvidia and Microsoft.

The AI company responsible for the Claude chatbot said the collaboration will help defend against rising AI-powered cyberattacks, particularly after it detected thousands of previously unknown bugs online.

Known as Project Glasswing, the initiative harnesses Anthropic’s new Mythos Preview frontier model, which the company said would only be made available to a handful of organisations. It will not be publicly available. Instead, a group of companies that are responsible for critical software infrastructure will be able to use Mythos Preview to secure their own software and open-source offerings.

Already, it has detected flaws in every major operating system and web browser – even finding one software bug that was 27 years old.

“Mythos Preview has already found thousands of high-severity vulnerabilities, including some in every major operating system and web browser,” the company said via a statement yesterday. “Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely.”

It added: “Project Glasswing is an urgent attempt to put these capabilities to work for defensive purposes.”

A previously unreleased AI model, Claude Mythos Preview is technology that uses strong agentic coding and reasoning skills to confront cybersecurity tasks, Anthropic said.

In a bid to accelerate its cybersecurity efforts, Anthropic said it will commit US$100 million worth of credits and $4 million in direct donations to open-source cybersecurity organisations.

“No one organisation can solve these cybersecurity problems alone: frontier AI developers, other software companies, security researchers, open-source maintainers and governments across the world all have essential roles to play,” the company added in its statement. “The work of defending the world’s cyber infrastructure might take years; frontier AI capabilities are likely to advance substantially over just the next few months.

“For cyber defenders to come out ahead, we need to act now.”

While working alongside big tech, Anthropic is also having discussions with the US government about Claude Mythos Preview and its potential ability to support defensive and offensive abilities. However, this is alongside the company’s current legal case with the Pentagon over its decision to have guidelines on how the US Department of Defense can use its AI models.

The news comes alongside a major Claude outage, with users seeing higher error rates than normal using Sonnet 4.6 this morning. It also comes one week after Anthropic accidentally leaked some of its internal source code, on account of human error.

Algorithms at war: How AI is shaping the conflict with Iran

Anthropic opens Bengaluru office to build strong AI partnerships