As the artificial intelligence space keeps growing into a larger behemoth every day, the giants are preparing for the next steps.
Anthropic is making an alliance with Nvidia, Microsoft and many other technology giants to deal with the next wave of AI systems, which may become powerful enough to reshape cyber defences.
Under a new initiative called Project Glasswing, the startup is giving selected partners access to an unreleased model, Claude Mythos Preview.
Anthropic said that the system has already discovered “thousands” of major software vulnerabilities.
Anthropic’s launch partners include Microsoft, Nvidia, Amazon, Apple, Google, CrowdStrike, Palo Alto Networks, Cisco, Broadcom, JPMorgan Chase and the Linux Foundation.
A model powerful enough to alarm its own maker
Anthropic’s basic argument is that Mythos Preview is not simply a better coding model.
It is unusually capable of finding hidden weaknesses in software that underpins the modern digital economy.
In its announcement, the company said the model identified thousands of significant vulnerabilities in operating systems.
That helps explain why Anthropic is not rolling it out like a normal product launch. Instead, it is treating the model as a tool that must first be tested under controlled conditions.
That distinction matters, as a model that can discover flaws quickly can also compress the time between identifying a vulnerability and exploiting it.
Anthropic’s public framing is that the safest first use case for such a system is defensive: helping trusted organizations find and fix weaknesses before attackers do.
Also read- Anthropic-Pentagon clash raises key question: who is to blame if AI kills?
Why Microsoft, Nvidia and others matter
The partner list shows that Anthropic is trying to build a defensive coalition around the parts of the digital system that matter most.
Microsoft brings reach across enterprise software, cloud computing and security tools. Nvidia sits at the center of the AI hardware stack.
Amazon, Google and Apple control large chunks of cloud, consumer and software infrastructure.
CrowdStrike, Palo Alto Networks and Cisco are deeply embedded in corporate cyber defense.
Broadcom, JPMorgan Chase and the Linux Foundation each represent critical layers of infrastructure, finance and open-source software.
In practical terms, these are the kinds of organizations that can spot threats early, test systems at scale and patch vulnerabilities quickly.
Anthropic said launch partners will use Mythos Preview in their defensive security work and that the company will share findings with the industry.
It also plans to expand access to about 40 additional organizations responsible for critical software infrastructure.
That makes Project Glasswing look less like a branding exercise and more like an attempt to put powerful AI tools first in the hands of institutions that can absorb the risks.
The deeper story is that AI is increasingly becoming both a threat and a defense.
This year’s RSA cybersecurity conference was dominated by discussion of AI-powered attacks and whether conventional defenses are still adequate.
The issue is not only that attackers can use AI to move faster, but that defenders may need equally capable systems just to keep up.
The post Why Anthropic is teaming up with Nvidia, Microsoft on cybersecurity appeared first on Invezz
